ブログ
/
Network
/
December 16, 2024

Cleo File Transfer Vulnerability: Patch Pitfalls and Darktrace’s Detection of Post-Exploitation Activities

File transfer applications are prime targets for ransomware groups due to their critical role in business operations. Recent vulnerabilities in Cleo's MFT software, namely CVE-2024-50623 and CVE-2024-55956, highlight ongoing risks. Read more about the Darktrace Threat Research team’s investigation into these vulnerabilities.
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
Maria Geronikolou
Cyber Analyst
Default blog image
16
Dec 2024

File transfer applications: A target for ransomware

File transfer applications have been a consistent target, particularly for ransomware groups, in recent years because they are key parts of business operations and have trusted access across different parts of an organization that include potentially confidential and personal information about an organization and its employees.

Recent targets of ransomware criminals includes applications like Acellion, Moveit, and GoAnywhere [1]. This seems to have been the case for Cleo’s managed file transfer (MFT) software solutions and the vulnerability CVE-2024-50623.

Threat overview: Understanding Cleo file transfer vulnerability

This vulnerability was believed to have been patched with the release of version 5.8.0.21 in late October 2024. However, open-source intelligence (OSINT) reported that the Clop ransomware group had managed to bypass the initial patch in late November, leading to the successful exploitation of the previously patched CVE.

In the last few days Cleo has published a new vulnerability, CVE-2024-55956, which is not a patch bypass of the CVE-2024-50623 but rather another vulnerability. This is also an unauthenticated file write vulnerability but while CVE-2024-50623 allows for both reading and writing arbitrary files, the CVE-2024-55956 only allows for writing arbitrary files and was addressed in version 5.8.0.24 [2].

Darktrace Threat Research analysts have already started investigating potential signs of devices running the Cleo software with network traffic supporting this initial hypothesis.

Comparison of CVE-2024-50623 and CVE-2024-55956

While CVE-2024-50623 was initially listed as a cross-site scripting issue, it was updated on December 10 to reflect unrestricted file upload and download. This vulnerability could lead to remote code execution (RCE) in versions of Cleo’s Harmony, VLTrader, and LexiCom products prior to 5.8.0.24. Attackers could leverage the fact that files are placed in the "autorun" sub-directory within the installation folder and are immediately read, interpreted, and evaluated by the susceptible software [3].

CVE-2024-55956, refers to an unauthenticated user who can import and execute arbitrary Bash or PowerShell commands on the host system by leveraging the default settings of the Autorun directory [4]. Both CVEs have occurred due to separate issues in the “/Synchronization” endpoint.

Investigating post exploitation patterns of activity on Cleo software

Proof of exploitation

Darktrace’s Threat Research analysts investigated multiple cases where devices identified as likely running Cleo software were detected engaging in unusual behavior. Analysts also attempted to identify any possible association between publicly available indicators of compromise (IoCs) and the exploitation of the vulnerability, using evidence of anomalous network traffic.

One case involved an Internet-facing device likely running Cleo VLTrader software (based on its hostname) reaching out to the 100% rare Lithuanian IP 181.214.147[.]164 · AS 15440 (UAB Baltnetos komunikacijos).

This activity occurred in the early hours of December 8 on the network of a customer in the energy sector. Darktrace detected a Cleo server transferring around over 500 MB of data over multiple SSL connections via port 443 to the Lithuanian IP. External research reported that this IP appears to be a callback IP observed in post-exploitation activity of vulnerable Cleo devices [3].

While this device was regularly observed sending data to external endpoints, this transfer represented a small increase in data sent to public IPs and coupled with the rarity of the destination, triggered a model alert as well as a Cyber AI Analyst Incident summarizing the transfer. Unfortunately, due to the encrypted connection no further analysis of the transmitted data was possible. However, due to the rarity of the activity, Darktrace’s Autonomous Response intervened and prevented any further connections to the IP.

 Model Alert Event Log show repeated connections to the rare IP, filtered with the rarity metric.
Figure 1: Model Alert Event Log show repeated connections to the rare IP, filtered with the rarity metric.
Shows connections to 181.214.147[.]164 and the amount of data transferred.
Figure 2: Shows connections to 181.214.147[.]164 and the amount of data transferred.

On the same day, external connections were observed to the external IP 45.182.189[.]225, along with inbound SSL connections from the same endpoint. OSINT has also linked this IP to the exploitation of Cleo software vulnerabilities [5].

Outgoing connections from a Cleo server to an anomalous endpoint.
Figure 3: Outgoing connections from a Cleo server to an anomalous endpoint.
 Incoming SSL connections from the external IP 45.182.189[.]225.
Figure 4: Incoming SSL connections from the external IP 45.182.189[.]225.

Hours after the last connection to 181.214.147[.]164, the integration detection tool from CrowdStrike, which the customer had integrated with Darktrace, issued an alert. This alert provided additional visibility into host-level processes and highlighted the following command executed on the Cleo server:

“D:\VLTrader\jre\bin\java.exe" -jar cleo.4889

Figure 5: The executed comand “D:\VLTrader\jre\bin\java.exe" -jar cleo.4889 and the Resource Location: \Device\HarddiskVolume3\VLTrader\jre\bin\java.exe.

Three days later, on December 11, another CrowdStrike integration alert was generated, this time following encoded PowerShell command activity on the server. This is consistent with post-exploitation activity where arbitrary PowerShell commands are executed on compromised systems leveraging the default settings of the Autorun directory, as highlighted by Cleo support [6]. According to external researchers , this process initiates connections to an external IP to retrieve JAR files with webshell-like functionality for continued post-exploitation [3]. The IP embedded in both commands observed by Darktrace was 38.180.242[.]122, hosted on ASN 58061(Scalaxy B.V.). There is no OSINT associating this IP with Cleo vulnerability exploitation at the time of writing.

Another device within the same customer network exhibited similar data transfer and command execution activity around the same time, suggesting it had also been compromised through this vulnerability. However, this second device contacted a different external IP, 5.45.74[.]137, hosted on AS 58061 (Scalaxy B.V.).

Like the first device, multiple connections to this IP were detected, with almost 600 MB of data transferred over the SSL protocol.

The Security Integration Detection Model that was triggered  and the PowerShell command observed
Figure 6: The Security Integration Detection Model that was triggered  and the PowerShell command observed
 Incoming connections from the external IP 38.180.242[.]122.
Figure 7: Incoming connections from the external IP 38.180.242[.]122.
Connections to the external IP 5.45.74[.]137.
Figure 8: Connections to the external IP 5.45.74[.]137.
Figure 9: Autonomous Response Actions triggered during the suspicious activities

While investigating potential Cleo servers involved in similar outgoing data activity, Darktrace’s Threat Research team identified two additional instances of likely Cleo vulnerability exploitation used to exfiltrate data outside the network. In those two instances, unusual outgoing data transfers were observed to the IP 176.123.4[.]22 (AS 200019, AlexHost SRL), with around 500 MB of data being exfiltrated over port 443 in one case (the exact volume could not be confirmed in the other instance). This IP was found embedded in encoded PowerShell commands examined by external researchers in the context of Cleo vulnerability exploitation investigations.

Conclusion

Overall, Cleo software represents a critical component of many business operations, being utilized by over 4,000 organizations worldwide. This renders the software an attractive target for threat actors who aim at exploiting internet-facing devices that could be used to compromise the software’s direct users but also other dependent industries resulting in supply chain attacks.

Darktrace / NETWORK was able to capture traffic linked to exploitation of CVE-2024-50623 within models that triggered such as Unusual Activity / Unusual External Data to New Endpoint while its Autonomous Response capability successfully blocked the anomalous connections and exfiltration attempts.

Information on new CVEs, how they're being exploited, and whether they've been patched can be fast-changing, sometimes limited and often confusing. Regardless, Darktrace is able to identify and alert to unusual behavior on these systems, indicating exploitation.

Credit to Maria Geronikolou, Alexandra Sentenac, Emma Fougler, Signe Zaharka and the Darktrace Threat Research team

[related-resource]

Appendices

References

[1] https://blog.httpcs.com/en/file-sharing-and-transfer-software-the-new-target-of-hackers/

[2] https://attackerkb.com/topics/geR0H8dgrE/cve-2024-55956/rapid7-analysis

[3] https://www.huntress.com/blog/threat-advisory-oh-no-cleo-cleo-software-actively-being-exploited-in-the-wild

[4] https://nvd.nist.gov/vuln/detail/CVE-2024-55956

[5] https://arcticwolf.com/resources/blog/cleopatras-shadow-a-mass-exploitation-campaign/

[6] https://support.cleo.com/hc/en-us/articles/28408134019735-Cleo-Product-Security-Advisory-CVE-Pending

[7] https://support.cleo.com/hc/en-us/articles/360034260293-Local-HTTP-Users-Configuration

Darktrace Model Alerts

Anomalous Connection / Data Sent to Rare Domain

Unusual Activity / Unusual External Data to New Endpoint

Unusual Activity / Unusual External Data Transfer

Device / Internet Facing Device with High Priority Alert

Anomalous Server Activity / Rare External from Server

Anomalous Connection / New User Agent to IP Without Hostname

Security Integration / High Severity Integration Incident

Security Integration / Low Severity Integration Detection

Autonomous Response Model Detections

Antigena / Network / Insider Threat / Antigena Large Data Volume Outbound Block

Antigena / Network / Significant Anomaly / Antigena Significant Server Anomaly Block

Antigena / Network / Significant Anomaly / Antigena Controlled and Model Alert

Cyber AI Analyst Incidents

Unusual External Data Transfer

MITRE ATT&CK Mapping

Tactic – Technique

INITIAL ACCESS – Exploit Public-Facing Application

COMMAND AND CONTROL – Application Layer Protocol (Web Protocols)

COMMAND AND CONTROL – Encrypted Channel

PERSISTENCE – Web Shell

EXFILTRATION - Exfiltration Over C2 Channel

IoC List

IoC       Type    Description + Probability

181.214.147[.]164      IP Address       Likely C2 Infrastructure

176.123.4[.]22            IP Address       Likely C2 Infrastructure

5.45.74[.]137               IP Address           Possible C2 Infrastructure

38.180.242[.]122        IP Address       Possible C2 Infrastructure

Get the latest insights on emerging cyber threats

Attackers are adapting, are you ready? This report explores the latest trends shaping the cybersecurity landscape and what defenders need to know.

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
Maria Geronikolou
Cyber Analyst

More in this series

No items found.

Blog

/

Proactive Security

/

June 2, 2026

Stopping Stealth Attacks with Precision: How Núclea Prevented a Breach Without Disruption

Default blog imageDefault blog image

Núclea is a Brazilian data and technology company that supports the country’s financial system by delivering digital services exclusively to banks and financial institutions. Operating in an environment where trust, availability, and data integrity are critical, the company faces a threat landscape that has evolved rapidly—particularly with the rise of AI-driven cyberattacks.

Brazil has experienced a wave of successful cyber incidents targeting financial institutions, many of them enabled by insiders or compromised credentials. The result was a noticeable shift in attacker strategy: instead of focusing on end customers, threat actors began targeting the institutions and platforms that underpin the financial ecosystem itself.

“Attacks became far more directed and contextual,” explains Guilherme, who leads incident response within Núclea’s security platform engineering team. “They weren’t noisy or obviously malicious—they were precise, patient, and designed to blend into normal operations.”

That precision was on full display in January 2026, when Núclea faced one of the most convincing phishing attacks the team had seen.

A real attack, built on trust and context

The attack began with a seemingly routine email.

It was sent from a real Brazilian government institution, using legitimate infrastructure and valid credentials that were later confirmed to have been compromised. Núclea had an established, ongoing relationship with this organization, and the email’s language, tone, and subject matter aligned perfectly with the type of communication the recipient team handled every day.

Attached to the email was a PDF document containing content that looked entirely legitimate.

The problem? A single URL embedded inside that PDF.

“The message itself was correct. The sender was real. The context was familiar. Even the document content made sense,” Guilherme explains. “There was just one small element that didn’t belong.”

That small detail was enough to initiate a full attack chain.

What the attackers were trying to do

If clicked, the URL would have downloaded a malicious payload designed to:

  • Collect information about the user and device
  • Identify where the system was located within the financial ecosystem
  • Install remote access tools to maintain control
  • Deploy an infostealer to extract sensitive data
  • Execute anti-forensic scripts to erase traces of the intrusion

In other words, it was a carefully engineered operation designed for persistence and stealth, not immediate disruption.

The attack also employed urgency—a classic social engineering technique. When the link didn’t open as expected, employees requested assistance from the security team, insisting the document was important and needed to be accessed quickly.

This is precisely the kind of scenario where traditional security tools struggle: almost everything about the interaction is legitimate.

Where Darktrace made the difference

Instead of blocking the entire message or relying on known indicators of compromise, Darktrace focused on behavioral context.

Darktrace recognized:

  • That the sending organization was normally trusted
  • That the communication pattern matched historical behavior
  • That the PDF content itself was not suspicious

But it also identified that the URL embedded within the document deviated from established behavioral patterns.

Rather than disrupting business operations, Darktrace took precise action: it rewrote the URL, preventing the malicious download while leaving the rest of the email untouched.

“When we analyzed it afterward, it became clear how dangerous the attack would have been,” says Guilherme. “But it never progressed—because Darktrace acted at exactly the right point.”

Subsequent forensic analysis confirmed the payload’s malicious intent. The attack never succeeded.

Precision over disruption

For Núclea, this incident reinforced a critical lesson: modern attacks don’t always look malicious—they hide within normal activity.

“What stands out to me is the precision,” Guilherme says. “Darktrace doesn’t rely on big, obvious signals. It’s effective in situations that fall outside the standard patterns we all know.”

Building resilience in a high trust ecosystem

For Núclea, cybersecurity is not just a defensive measure—it’s a business enabler.

Availability failures or successful breaches in the financial ecosystem can have immediate, large-scale consequences, from financial loss to reputational damage. Preventing those outcomes protects not just Núclea, but its partners and customers as well.

“Cyber resilience means keeping the business running—even under attack,” Guilherme explains. “And that requires people, processes, and technology working together.”

As AI continues to accelerate both attacks and defenses, the role of security is evolving. Precision, behavioral understanding, and intelligent automation are no longer optional—they’re essential.

“The easy days were yesterday,” Guilherme says. “The challenges ahead are bigger. We need to be prepared—internally and with partners that help us build resilience.”

Continue reading
About the author

Blog

/

AI

/

June 1, 2026

効率化の裏にあるリスク:AI導入が製造現場にもたらす見えない脆弱性

Default blog imageDefault blog image

AIエージェントが製造業に与える影響

製造業界のセキュリティチームやIT担当者は、生産を守り、稼働時間を維持し、重要資産を保護するという絶え間ないプレッシャー下にあります。そしてAIは非常に大きなチャンスとともに、新たなサイバーリスクももたらしています。製造業全体で、AIはワークフローや意思決定に組み込まれつつあり、自律型AIエージェントが従業員やシステムに代わって行動する場面が増えています。

エージェント型システムは独立して行動できるため強力ですが、その同じ自律性がサイバーリスク、運用上のリスクも生み出します。エージェントは広範な権限を持ち、複雑なタスクの実行、意思決定、ツールや外部システムとのやり取りを、ほとんどまたは全く人間の介入なしに行うことができます。

あらかじめ定義されたタスクを実行する従来のAIモデルとは異なり、AIエージェントは高度なテクニックを使用して人間の意思決定プロセスを模倣することにより、新たな課題に動的に適応し、また自らの判断に基づいて意思決定し、アクションを実行します。彼らは業務の上では従業員のように見えますが、人間が持つ判断力、倫理観、または行動の結果に対する恐れが欠けています。これは、サイバー犯罪者によって簡単に操られる可能性があることを意味しており、OTネットワーク全体に埋め込まれたAIエージェントは、データ漏洩をはるかに超える脅威を生み出します。たとえば、BMWでは、AI は溶接プロセスのエラーの発生を識別するのに使われています。同社のスパータンバーグ(米サウスカロライナ州)の工場では、すべてのSUVフレーム上の300-400個のスタッドの溶接をAIが監視し、スタッドの配置間違いや欠陥を検知し直ちに修正します。このAIシステムが破損すれば壊滅的な品質管理問題につながる恐れがあります。

製造全体にエージェント型AIシステムを導入することについて多くのセキュリティチームはさまざまな懸念を示しています。ダークトレースの行ったAIサイバーセキュリティの現状調査では、製造業のセキュリティプロフェッショナルの78%が従業員によるAIエージェントの利用に懸念を抱いており、これは彼らの最も大きな危惧でした。それに続く問題点が従業員によるCopilotやChatGPT等の生成AIツールの使用であり、製造業のセキュリティプロフェッショナルの76%が懸念を抱いていました。これらのツールがますます多くのビジネスデータやプロセスにアクセスし、組織内でより多くの自律性を持つようになるにつれ、エージェントのアクティビティがほとんど可視化されていない現在、セキュリティチームにおいては機密データの露出(60%)や偶発的なポリシーおよび規制違反(59%)への懸念が高まっています。

外部からのAIによる脅威も急激に進化

製造業を変革しているのと同じAIの能力が、サイバー攻撃の形も変貌させています。

AIにより攻撃者は偵察を自動化し、標的をより高度に絞り込み、リアルタイムで適応できるようになっています。かつては人手による作業と時間を要していたことが、今では継続的かつ大規模に実行できるようになりました。そして、製造業はすでにその影響を実感しています。当社が調査した製造業のセキュリティプロフェッショナルの76%は、すでにAIを活用した脅威の影響を受けており、90%がAIによってソーシャルエンジニアリング攻撃の成功率が高まっていると回答しています。

また、攻撃のテクニック自体も進化しています。製造業界全体で、AIを利用した攻撃の経路の多様化に対する懸念が高まっています。特にリアルタイムで進化する適応型マルウェアについて、調査対象の製造業のセキュリティプロフェッショナルの半数近く(49%)が懸念しており、これは全産業の平均よりも9%高い数値です。AIを使った適応型マルウェアに続くその他の懸念には次が含まれます:

  • 自動化された脆弱性スキャンとエクスプロイトチェイニング(48%):Anthropicの新しいMythos AIモデルにより脆弱性探索が深刻化する中で、この問題は一層差し迫ったものとなっています。
  • 超パーソナライズされたフィッシングキャンペーン(46%):フィッシングは依然としてハッカーの主力兵器の1つであり、AIによってフィッシングメールはより説得力が高く検知困難なものとなり、その効果は増幅されました。

これは単に攻撃の量の増加だけでなく、攻撃の展開につれて静的な防御が対応できるよりも速く進化する脅威への変化なのです。

こうした認識が高まっているにもかかわらず、製造業の多くはまだこの変化に対応する準備ができていません。半数以上(51%)がAI駆動の脅威への準備が十分にできていないと回答し、AIの導入を管理する正式なポリシーを持っている組織はわずか37%でした。  

可視性、コンテキスト、およびガードレールを通じてAIのセキュリティを確保

これらの問題に対処するためにAIイノベーションを遅らせる必要はありません。それには、AIと同じスピードと規模で動作できる、これまでとは異なるアプローチのセキュリティが必要です。具体的には、製造業がAIの力を活用する上で、次の3つの優先課題が浮上しています。

可視性はすべての土台  

AIがどこで使用されているか、何にアクセスできるか、そしてITおよびOT環境にわたってどのように動作するかを理解する必要があります。それがなければ、リスクを測定したり管理したりすることはできません。ダークトレースの調査において、製造業のセキュリティプロフェッショナルの91%が、AIを信頼する前に、それがどのように意思決定を行うかを理解する必要があると回答したのは当然のことです。OT環境においてこのことはさらに重要です。稼働の中断は安全や環境、財務、および評判に大きな影響を及ぼすからです。

可視性をアクションにつなげるにはコンテキストが必要  

AIによって形作られる環境において、正常とされる挙動は絶えず変化します。つまり、脅威を検知するにはビヘイビアベースのアプローチが必要なのです。組織全体で生活パターンを理解し、わずかな逸脱をリアルタイムに検知すること- これは従来のセキュリティとリスク管理に対するアプローチからの根本的な変化です。

エージェントからの露出を防ぐガードレール  

AIシステムがより大きな責任を担うようになるなかで、組織はAIが何をできるか、そしていつ独立して行動できるかについて、明確な境界を設ける必要があります。これらのコントロールは何かがあってから適用されるのではなく、システム自体に組み込んでおかなければなりません。  

製造業のITおよびOT環境におけるAIエージェントのセキュリティ

エージェント型AIの出現は製造業を変革し、次世代のオペレーションを支える一方で、脅威ランドスケープも一変させています。これは単なる脅威の増加ではなく、自律型システムへの移行、挙動の絶え間ない変化、そしてマシンスピードで進行するリスクです。AIを活用しつつリスクを管理するという課題に取り組む組織にとって、可視性、コンテキスト、ガードレールはセキュリティの基盤となります。

Darktraceはこの基盤を実現することにより、製造業の安全なAIアプローチ構築を支援します。ITおよびOT環境全体を可視化し、異常なアクティビティに対するリアルタイムの検知および対応を提供することにより、従業員が使用するプロンプトや構築するエージェントから、それらのエージェントの環境全体での動作に至るまで、AIアクティビティの理解を可能にします。これにより、AIの導入を拡大する製造業はコントロールを犠牲にすることなくイノベーションの基盤を構築することができます。

Continue reading
About the author
Dr. Oakley Cox-Robinson
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