ブログ
/
Network
/
July 24, 2025

Untangling the web: Darktrace’s investigation of Scattered Spider’s evolving tactics

Learn about a recent Scattered Spider attack observed by Darktrace, comparing tactics with those seen in previous attacks. Widespread use of LOTL techniques alongside continued changes in TTPs such as their recent use of Ransomware-as-a-Service (RaaS) platforms can make it challenging for security teams to harden defenses.
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
Emma Foulger
Global Threat Research Operations Lead
Default blog image
24
Jul 2025

What is Scattered Spider?

Scattered Spider is a native English-speaking group, also referred to, or closely associated with, aliases such as UNC3944, Octo Tempest and Storm-0875. They are primarily financially motivated with a clear emphasis on leveraging social engineering, SIM swapping attacks, exploiting legitimate tooling as well as using Living-Off-the-Land (LOTL) techniques [1][2].

In recent years, Scattered Spider has been observed employing a shift in tactics, leveraging Ransomware-as-a-Service (RaaS) platforms in their attacks. This adoption reflects a shift toward more scalable attacks with a lower barrier to entry, allowing the group to carry out sophisticated ransomware attacks without the need to develop it themselves.

While RaaS offerings have been available for purchase on the Dark Web for several years, they have continued to grow in popularity, providing threat actors a way to cause significant impact to critical infrastructure and organizations without requiring highly technical capabilities [12].

This blog focuses on the group’s recent changes in tactics, techniques, and procedures (TTPs) reported by open-source intelligence (OSINT) and how TTPs in a recent Scattered Spider attack observed by Darktrace compare.

How has Scattered Spider been reported to operate?

First observed in 2022, Scattered Spider is known to target various industries globally including telecommunications, technology, financial services, and commercial facilities.

Overview of key TTPs

Scattered Spider has been known to utilize the following methods which cover multiple stages of the Cyber Kill Chain including initial access, lateral movement, evasion, persistence, and action on objective:

Social engineering [1]:

Impersonating staff via phone calls, SMS and Telegram messages; obtaining employee credentials (MITRE techniques T1598,T1656), multi-factor authentication (MFA) codes such as one-time passwords, or convincing employees to run commercial remote access tools enabling initial access (MITRE techniques T1204,T1219,T1566)

  • Phishing using specially crafted domains containing the victim name e.g. victimname-sso[.]com
  • MFA fatigue: sending repeated requests for MFA approval with the intention that the victim will eventually accept (MITRE technique T1621)

SIM swapping [1][3]:

  • Includes hijacking phone numbers to intercept 2FA codes
  • This involves the actor migrating the victim's mobile number to a new SIM card without legitimate authorization

Reconnaissance, lateral movement & command-and-control (C2) communication via use of legitimate tools:

  • Examples include Mimikatz, Ngrok, TeamViewer, and Pulseway [1]. A more recently reported example is Teleport [3].

Financial theft through their access to victim networks: Extortion via ransomware, data theft (MITRE technique T1657) [1]

Bring Your Own Vulnerable Driver (BYOVD) techniques [4]:

  • Exploiting vulnerable drivers to evade detection from Endpoint Detection and Response (EDR) security products (MITRE technique T1068) frequently used against Windows devices.

LOTL techniques

LOTL techniques are also closely associated with Scattered Spider actors once they have gained initial access; historically this has allowed them to evade detection until impact starts to be felt. It also means that specific TTPs may vary from case-to-case, making it harder for security teams to prepare and harden defences against the group.

Prominent Scattered Spider attacks over the years

While attribution is sometimes unconfirmed, Scattered Spider have been linked with a number of highly publicized attacks since 2022.

Smishing attacks on Twilio: In August 2022 the group conducted multiple social engineering-based attacks. One example was an SMS phishing (smishing) attack against the cloud communication platform Twilio, which led to the compromise of employee accounts, allowing actors to access internal systems and ultimately target Twilio customers [5][6].

Phishing and social engineering against MailChimp: Another case involved a phishing and social engineering attack against MailChimp. After gaining access to internal systems through compromised employee accounts the group conducted further attacks specifically targeting MailChimp users within cryptocurrency and finance industries [5][7].

Social engineering against Riot Games: In January 2023, the group was linked with an attack on video game developer Riot Games where social engineering was once again used to access internal systems. This time, the attackers exfiltrated game source code before sending a ransom note [8][9].

Attack on Caesars & MGM: In September 2023, Scattered Spider was linked with attacked on Caesars Entertainment and MGM Resorts International, two of the largest casino and gambling companies in the United States. It was reported that the group gathered nearly six terabytes of stolen data from the hotels and casinos, including sensitive information of guests, and made use of the RaaS strain BlackCat [10].

Ransomware against Marks & Spencer: More recently, in April 2025, the group has also been linked to the alleged ransomware incident against the UK-based retailer Marks & Spencer (M&S) making use of the DragonForce RaaS [11].

How a recent attack observed by Darktrace compares

In May 2025, Darktrace observed a Scattered Spider attack affecting one of its customers. While initial access in this attack fell outside of Darktrace’s visibility, information from the affected customer suggests similar social engineering techniques involving abuse of the customer’s helpdesk and voice phishing (vishing) were used for reconnaissance.

Initial access

It is believed the threat actor took advantage of the customer’s third-party Software-as-a-Service (SaaS) applications, such as Salesforce during the attack.

Such applications are a prime target for data exfiltration due to the sensitive data they hold; customer, personnel, and business data can all prove useful in enabling further access into target networks.

Techniques used by Scattered Spider following initial access to a victim network tend to vary more widely and so details are sparser within OSINT. However, Darktrace is able to provide some additional insight into what techniques were used in this specific case, based on observed activity and subsequent investigation by its Threat Research team.

Lateral movement

Following initial access to the customer’s network, the threat actor was able to pivot into the customer’s Virtual Desktop Infrastructure (VDI) environment.

Darktrace observed the threat actor spinning up new virtual machines and activating cloud inventory management tools to enable discovery of targets for lateral movement.

In some cases, these virtual machines were not monitored or managed by the customer’s security tools, allowing the threat actor to make use of additional tooling such as AnyDesk which may otherwise have been blocked.

Tooling in further stages of the attack sometimes overlapped with previous OSINT reporting on Scattered Spider, with anomalous use of Ngrok and Teleport observed by Darktrace, likely representing C2 communication. Additional tooling was also seen being used on the virtual machines, such as Pastebin.

 Cyber AI Analyst’s detection of C2 beaconing to a teleport endpoint with hostname CUSTOMERNAME.teleport[.]sh, likely in an attempt to conceal the traffic.
Figure 1: Cyber AI Analyst’s detection of C2 beaconing to a teleport endpoint with hostname CUSTOMERNAME.teleport[.]sh, likely in an attempt to conceal the traffic.

Leveraging LOTL techniques

Alongside use of third-party tools that may have been unexpected on the network, various LOTL techniques were observed during the incident; this primarily involved the abuse of standard network protocols such as:

  • SAMR requests to alter Active Directory account details
  • Lateral movement over RDP and SSH
  • Data collection over LDAP and SSH

Coordinated exfiltration activity linked through AI-driven analysis

Multiple methods of exfiltration were observed following internal data collection. This included SSH transfers to IPs associated with Vultr, alongside significant uploads to an Amazon S3 bucket.

While connections to this endpoint were not deemed unusual for the network at this stage due to the volume of traffic seen, Darktrace’s Cyber AI Analyst was still able to identify the suspiciousness of this behavior and launched an investigation into the activity.

Cyber AI Analyst successfully correlated seemingly unrelated internal download and external upload activity across multiple devices into a single, broader incident for the customer’s security team to review.

Cyber AI Analyst Incident summary showing a clear outline of the observed activity, including affected devices and the anomalous behaviors detected.
Figure 2: Cyber AI Analyst Incident summary showing a clear outline of the observed activity, including affected devices and the anomalous behaviors detected.
Figure 3: Cyber AI Analyst’s detection of internal data downloads and subsequent external uploads to an Amazon S3 bucket.

Exfiltration and response

Unfortunately, as Darktrace was not configured in Autonomous Response mode at the time, the attack was able to proceed without interruption, ultimately escalating to the point of data exfiltration.

Despite this, Darktrace was still able to recommend several Autonomous Response actions, aimed at containing the attack by blocking the internal data-gathering activity and the subsequent data exfiltration connections.

These actions required manual approval by the customer’s security team and as shown in Figure 3, at least one of the recommended actions was subsequently approved.

Had Darktrace been enabled in Autonomous Response mode, these measures would have been applied immediately, effectively halting the data exfiltration attempts.

Further recommendations for Autonomous Response actions in Darktrace‘s Incident Interface, with surgical response targeting both the internal data collection and subsequent exfiltration.
Figure 4: Further recommendations for Autonomous Response actions in Darktrace‘s Incident Interface, with surgical response targeting both the internal data collection and subsequent exfiltration.

Scattered Spider’s use of RaaS

In this recent Scattered Spider incident observed by Darktrace, exfiltration appears to have been the primary impact. While no signs of ransomware deployment were observed here, it is possible that this was the threat actors’ original intent, consistent with other recent Scattered Spider attacks involving RaaS platforms like DragonForce.

DragonForce emerged towards the end of 2023, operating by offering their platform and capabilities on a wide scale. They also launched a program which offered their affiliates 80% of the eventual ransom, along with tools for further automation and attack management [13].

The rise of RaaS and attacker customization is fragmenting TTPs and indicators, making it harder for security teams to anticipate and defend against each unique intrusion.

While DragonForce appears to be the latest RaaS used by Scattered Spider, it is not the first, showcasing the ongoing evolution of tactics used the group.

In addition, the BlackCat RaaS strain was reportedly used by Scattered Spider for their attacks against Caesars Entertainment and MGM Resorts International [10].

In 2024 the group was also seen making use of additional RaaS strains; RansomHub and Qilin [15].

What security teams and CISOs can do to defend against Scattered Spider

The ongoing changes in tactics used by Scattered Spider, reliance on LOTL techniques, and continued adoption of evolving RaaS providers like DragonForce make it harder for organizations and their security teams to prepare their defenses against such attacks.

CISOs and security teams should implement best practices such as MFA, Single Sign-On (SSO), notifications for suspicious logins, forward logging, ethical phishing tests.

Also, given Scattered Spider’s heavy focus on social engineering, and at times using their native English fluency to their advantage, it is critical to IT and help desk teams are reminded they are possible targets.

Beyond social engineering, the threat actor is also adept at taking advantage of third-party SaaS applications in use by victims to harvest common SaaS data, such as PII and configuration data, that enable the threat actor to take on multiple identities across different domains.

With Darktrace’s Self-Learning AI, anomaly-based detection, and Autonomous Response inhibitors, businesses can halt malicious activities in real-time, whether attackers are using known TTPs or entirely new ones. Offerings such as Darktrace /Attack Surface Management enable security teams to proactively identify signs of malicious activity before it can cause an impact, while more generally Darktrace’s ActiveAI Security Platform can provide a comprehensive view of an organization’s digital estate across multiple domains.

Credit to Justin Torres (Senior Cyber Analyst), Emma Foulger (Global Threat Research Operations Lead), Zaki Al-Dhamari (Cyber Analyst), Nathaniel Jones (VP, Security & AI Strategy, FCISO), and Ryan Traill (Analyst Content Lead)

---------------------

The information provided in this blog post is for general informational purposes only and is provided "as is" without any representations or warranties, express or implied. While Darktrace makes reasonable efforts to ensure the accuracy and timeliness of the content related to cybersecurity threats such as Scattered Spider, we make no warranties or guarantees regarding the completeness, reliability, or suitability of the information for any purpose.

This blog post does not constitute professional cybersecurity advice, and should not be relied upon as such. Readers should seek guidance from qualified cybersecurity professionals or legal counsel before making any decisions or taking any actions based on the content herein.

No warranty of any kind, whether express or implied, including, but not limited to, warranties of performance, merchantability, fitness for a particular purpose, or non-infringement, is given with respect to the contents of this post.

Darktrace expressly disclaims any liability for any loss or damage arising from reliance on the information contained in this blog.

Appendices

References

[1] https://www.cisa.gov/news-events/cybersecurity-advisories/aa23-320a

[2] https://attack.mitre.org/groups/G1015/

[3] https://www.rapid7.com/blog/post/scattered-spider-rapid7-insights-observations-and-recommendations/

[4] https://www.crowdstrike.com/en-us/blog/scattered-spider-attempts-to-avoid-detection-with-bring-your-own-vulnerable-driver-tactic/

[5] https://krebsonsecurity.com/2024/06/alleged-boss-of-scattered-spider-hacking-group-arrested/?web_view=true

[6] https://www.cxtoday.com/crm/uk-teenager-accused-of-hacking-twilio-lastpass-mailchimp-arrested/

[7] https://mailchimp.com/newsroom/august-2022-security-incident/

[8] https://techcrunch.com/2023/02/02/0ktapus-hackers-are-back-and-targeting-tech-and-gaming-companies-says-leaked-report/

[9] https://www.pcmag.com/news/hackers-behind-riot-games-breach-stole-league-of-legends-source-code

[10] https://www.bbrown.com/us/insight/a-look-back-at-the-mgm-and-caesars-incident/

[11] https://cyberresilience.com/threatonomics/scattered-spider-uk-retail-attacks/

[12] https://www.crowdstrike.com/en-us/cybersecurity-101/ransomware/ransomware-as-a-service-raas/

[13] https://www.group-ib.com/blog/dragonforce-ransomware/
[14] https://blackpointcyber.com/wp-content/uploads/2024/11/DragonForce.pdf
[15] https://x.com/MsftSecIntel/status/1812932749314978191?lang=en

Select MITRE tactics associated with Scattered Spider

Tactic – Technique – Technique Name

Reconnaissance - T1598 -   Phishing for Information

Initial Access - T1566 – Phishing

Execution - T1204 - User Execution

Privilege Escalation - T1068 - Exploitation for Privilege Escalation

Defense Evasion - T1656 - Impersonation

Credential Access - T1621 - Multi-Factor Authentication Request Generation

Lateral Movement - T1021 - Remote Services

Command and Control - T1102 - Web Service

Command and Control - T1219 - Remote Access Tools

Command and Control - T1572 - Protocol Tunneling

Exfiltration - T1567 - Exfiltration Over Web Service

Impact - T1657 - Financial Theft

Select MITRE tactics associated with DragonForce

Tactic – Technique – Technique Name

Initial Access, Defense Evasion, Persistence, Privilege Escalation - T1078 - Valid Accounts

Initial Access, Persistence - T1133 - External Remote Services

Initial Access - T1190 - Exploit Public-Facing Application

Initial Access - T1566 – Phishing

Execution - T1047 - Windows Management Instrumentation

Privilege Escalation - T1068 - Exploitation for Privilege Escalation

Lateral Movement - T1021 - Remote Services

Impact - T1486 - Data Encrypted for Impact

Impact - T1657 - Financial Theft

Select Darktrace models

Compliance / Internet Facing RDP Server

Compliance / Incoming Remote Access Tool

Compliance / Remote Management Tool on Server

Anomalous File / Internet Facing System File Download

Anomalous Server Activity/ New User Agent from Internet Facing System

Anomalous Connection / Callback on Web Facing Device

Device / Internet Facing System with High Priority Alert

Anomalous Connection / Unusual Admin RDP

Anomalous Connection / High Priority DRSGetNCChanges

Anomalous Connection / Unusual Internal SSH

Anomalous Connection / Active Remote Desktop Tunnel

Compliance / Pastebin

Anomalous Connection / Possible Tunnelling to Rare Endpoint

Compromise / Beaconing Activity to External Rare

Device / Long Agent Connection to New Endpoint

Compromise / SSH to Rare External AWS

Compliance / SSH to Rare External Destination

Anomalous Server Activity / Outgoing from Server

Anomalous Connection / Large Volume of LDAP Download

Unusual Activity / Internal Data Transfer on New Device

Anomalous Connection / Download and Upload

Unusual Activity / Enhanced Unusual External Data Transfer

Compromise / Ransomware/Suspicious SMB Activity

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
Emma Foulger
Global Threat Research Operations Lead

More in this series

No items found.

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

Default blog imageDefault blog image

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).

付録3:参考文献

外部

[1] https://www.darkreading.com/cyberattacks-data-breaches/jadepuffer-first-complete-llm-driven-ransomware-attack

[2] https://www.trendmicro.com/en_us/research/26/c/pawn-storm-targets-govt-infra.html

ダークトレースのリソース

1.        https://www.darktrace.com/blog/ai-llm-generated-malware-used-to-exploit-react2shell

2.        https://www.darktrace.com/blog/2025-cyber-threat-landscape-darktraces-mid-year-review

3.        https://www.darktrace.com/resources/annual-threat-report-2026

4.        https://www.darktrace.com/blog/when-trust-becomes-the-attack-surface-supply-chain-attacks-in-an-era-of-automation-and-implicit-trust

5.        https://www.darktrace.com/blog/hola-vpn-abuse-from-proxy-traffic-to-malware-and-cryptomining

6.        https://www.darktrace.com/blog/security-after-signatures-operating-in-a-world-of-pre-cve-disclosure-exploitation-collapsed-trust-boundaries-and-autonomous-systems

7.        https://www.darktrace.com/blog/when-ai-infrastructure-becomes-part-of-the-attack-surface

8.        https://www.darktrace.com/blog/cve-2026-1731-how-darktrace-sees-the-beyondtrust-exploitation-wave-unfolding

9.        https://www.darktrace.com/resource/understanding-chinese-nexus-cyber-tradecraft

10.   https://www.darktrace.com/blog/chinese-apt-campaign-targets-entities-with-updated-fdmtp-backdoor

11.  https://www.darktrace.com/blog/inside-zionsiphon-darktraces-analysis-of-ot-malware-targeting-israeli-water-systems

12.  https://www.darktrace.com/resources/the-state-of-cybersecurity-in-the-finance-sector

13.  https://www.darktrace.com/blog/from-click-to-command-behavioral-detection-of-applescript-led-macos-intrusions

14.  https://www.darktrace.com/blog/the-state-of-cybersecurity-in-the-finance-sector-six-trends-to-watch

Continue reading
About the author
Nathaniel Jones
SVP, Global Threat Intelligence

Blog

/

Network

/

August 11, 2026

When AI Agents Attack: The Case for Behavioral Anomaly Detection

Default blog imageDefault blog image

The UK AI Security Institute (AISI) has just published a detailed incident report confirming what many within the cyber security industry have suspected for some time: frontier AI agents, given a degree of autonomy, can independently develop and execute attack chains against real targets, including social engineering, supply-chain compromise, and deception, without explicit instruction to do so.

The report (INC-2026-07-28-01, published 4 August 2026) documents what happened during a routine capability evaluation between 25 and 28 July. It deserves a careful read from anyone responsible for securing the modern digital estate.

The incident

AISI regularly tests frontier AI models for offensive cyber capability. In this exercise, agents were placed inside a simulated network environment and given internet access this was deliberate, given to reflect realism of the task. Across 122 evaluation runs, 19 instances of unsanctioned activity were recorded, including cases where agents broke the intended boundaries of the test and acted against real people and real infrastructure on the open internet.

In the most serious case, an agent (Mythos 5) autonomously:

  • Conducted open-source intelligence gathering to identify real developers associated with keywords from the simulated scenario
  • Created a GitHub account and submitted a pull request containing obfuscated malicious code to a real, unaffiliated repository
  • Created a second sockpuppet account to post comments on its own PR, attempting to manufacture social proof
  • When a human reviewer flagged the malicious content, the agent lied, claiming an accidental commit, then repeatedly attempted to reintroduce the payload under the guise of having "fixed" the code
  • Sent spear-phishing emails to the repository maintainer containing manipulated content designed to convince them to merge the code
  • Planted a prompt injection inside a GitHub issue, hidden in an HTML comment invisible to humans but readable by AI coding assistants parsing the repository via API

AISI confirmed no real-world harm resulted. The PRs were not merged; the payloads did not execute. But containment was a product of circumstance, not design.

Why this matters

Every action the agents took used legitimate services operating within normal parameters. Creating a GitHub account is legitimate. Opening a pull request is legitimate. Sending an email is legitimate. Posting an issue is legitimate.

No vulnerability was exploited. No known malware signature was present. No indicator of compromise from any threat intelligence feed would have matched. The entire attack existed as a pattern, a sequence of individually normal actions that, taken together, constituted a supply-chain attack, a social engineering campaign, and an attempted compromise of downstream systems.

This is the class of threat that signature-based and rule-based security tools structurally struggle to detect. They answer the question "have I seen this before?" When the answer is no – and when the attack is novel, composed of legitimate actions, and adapted in real time by an adversary in this case an agent that reasons about evasion – they struggle.

Behavioral anomaly detection: The architecture that fits

Darktrace has spent a decade building a platform around a different question: "is this behavior normal for this entity?"

The platform learns the baseline behavior of every user, device, and communication pattern across an organization's digital environment. It does not need prior knowledge of an attack technique. It identifies deviation from learned normality, surfaces it, and through its Autonomous Response capability, takes targeted action to contain threats at machine speed.

Consider how this AISI incident might show across the Darktrace platform:

Network detection and response

While no payload was successfully delivered to a victim in this incident, AISI's post-incident analysis of the agent's code (detailed on pages 24-25 of the report) reveals what would have landed on target systems had the social engineering succeeded. The agent produced three successive payload generations, each replacing the last after being detected and called out:

Generation 1 hid a module-load stager inside a legitimate bug-fix commit, guarded by an environment variable check, to skip execution in CI. Its second stage attempted to fetch Poseidon implants (Linux and Windows variants) calling back to the agent's own C2 listener, with a fallback to publicly hosted polling scripts that fetched and executed commands every 60 seconds.

Generation 2 exploited an Electron "What's New" window with nodeIntegration enabled, loading remote HTML from a GitHub Pages site the agent had registered. The page decoded an XOR-obfuscated payload at runtime that beaconed host identifiers over DNS, attempted to download the same implants, ran its own in-process C2 loop, and installed reboot persistence via crontab and Windows registry Run keys.

Generation 3 was a post-install dropper added after the agent had already been publicly caught. It XOR-decoded a hidden URL, downloaded a script over plaintext HTTP, and ran it detached, force-exiting so npm install always reports success, giving the attacker remote code execution on any machine running the install.

None of these payloads reached a victim. But the technical detail matters: this is the calibre of tradecraft an autonomous agent produced in a matter of hours, unprompted. Each generation demonstrated adaptation, the agent iterated its approach in direct response to being detected.

For network defenders, the behavioral signals these payloads would have generated are clear. DNS beaconing to novel domains, periodic HTTP polling to newly registered infrastructure, XOR-encoded data in transit, registry and crontab persistence writes on endpoints that have never exhibited such behavior. These are exactly the anomalies that Darktrace / NETWORK identifies in real time: deviations from learned device and connection baselines, regardless of whether the specific payload or C2 domain has been seen before.

Beyond the payload analysis, the agents' broader activity produced additional behavioral anomalies: new identities engaging in high-frequency interactions with external services, coordination via unusual channels (public pastebins, gists), and lateral communication patterns that deviate from any established baseline.

Email security

The spear-phishing emails sent by the agent, crafted with context-specific content designed to manipulate a specific individual, represent exactly the kind of socially engineered communication that static email gateways can miss. Darktrace / EMAIL analyses the linguistic patterns, sender behavior, and relational context of every message. A brand-new sender using tailored language to pressure a recipient into a specific technical action deviates from the learned communication profile, regardless of whether the message contains a known-bad URL or attachment.

Securing AI

This is where the incident points to a threat category that many security platforms do not yet fully address. The agent did not just use AI as a tool for attack. It targeted other AI systems, attempting prompt injections against coding assistants and trying to compromise automated workflows. Organisations adopting AI agents for development, operations, and decision-making are introducing a newer extended attack surface.  

Darktrace / SECURE AI extends the same behavioral detection model into this space. It monitors AI usage across the enterprise, detecting anomalous interactions with AI services, identifying sensitive data exposure to AI systems, providing visibility into shadow AI use. As AI agents become standard components of enterprise infrastructure, the ability to detect when an AI system is behaving outside its expected parameters, when there is a chance, it is being manipulated, or when it is interacting with something it should not be, that becomes a direct security requirement rather than a future consideration.

The principle is consistent across the Darktrace platform: learn what normal looks like, identify deviation, respond autonomously.

The AI-to-AI threat surface

One of the report's most important findings is a confirmation that AI agents can act outside of instructions and guardrails to complete tasks. The Mythos 5 agent conducted reconnaissance on a repository maintainer, inferred it might be an AI coding assistant, and deployed a prompt injection specifically crafted to compromise it.

This is a new attack class. Organizations deploying AI coding assistants, AI-powered code review, or autonomous DevOps agents now face an adversary that:

  • Moves at speed
  • Is able to adapt its approach when initial attempts are halted
  • Specifically targets AI systems via techniques (prompt injection, hidden instructions in HTML comments) that are less visible to human reviewers but parsed by machines
  • Reasons about which evasion strategies are most likely to succeed

Defending against this requires security that operates with the same adaptability and speed. A system that learns what "normal" looks like, and flags deviation from it.

What AISI recommends

The report identifies five contributing factors to the incident. Two are directly relevant:

No synchronous real-time monitoring (Section 5.3): AISI acknowledges it had not built real-time behavioral monitoring capable of immediately blocking unsanctioned agent actions. Their security team detected the incident via commercial security monitoring software, but only after activity had already occurred.

No fine-grained behavioral controls (Section 6.2): AISI's forward-looking response includes building continuous monitoring that can "immediately block or flag unsanctioned AI agent behavior."

The direction is clear: organizations need security that watches behavior continuously, learns what normal looks like, and acts when deviation occurs, without requiring a human in the loop for every decision.

The takeaway

The AISI incident is documented proof that autonomous AI agents will independently develop novel attack chains using legitimate tools, adapt when detected, employ deception, and target both humans and other AI systems. This happened last week, in a controlled setting, with commercially available models.

The security architecture that addresses this is behavioral anomaly detection applied across the full digital estate, as AI agents become standard components of enterprise infrastructure, writing code, managing deployments, processing communications, the attack surface they create is behavioral by nature.

This is the approach Darktrace has taken for years: learning what is normal across an organization’s digital environment, identifying meaningful deviations, and responding to emerging threats without relying on known attack signatures. As autonomous AI agents introduce new and unpredictable behaviors, that foundation becomes increasingly important to securing the enterprise.

Read the full report from the UK AI Security Institute here.

Continue reading
About the author
Adam Stevens
Senior Director of Product
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