Join our Newsletter — 33% off our NHI Course
Home› Guides› APT37 Threat Actor Profile Guide: Defector targeting, ROKRAT…
Guide Threat Intelligence & Threat Actors

APT37 Threat Actor Profile Guide: Defector targeting, ROKRAT and defences explained

PDF for newsletter subscribers · free to read online

← All guides
By NHI Mgmt Group, edited by Lalit Choda Updated 27 October 2026 11 min read
On this page

APT37, also called ScarCruft, is a North Korean espionage group that targets South Korean government contacts, North Korean defectors, human rights activists and journalists through booby-trapped documents, hijacked messenger and email accounts, and browser zero-days, then watches them with tools such as ROKRAT. MITRE ATT&CK describes it as state-sponsored and active since at least 2012, and Google reported in December 2022 that it exploited an Internet Explorer zero-day using a lure about the Itaewon crowd disaster. This profile covers its names, targeting, the techniques MITRE ATT&CK records for it, how it abuses identities and credentials, notable reported campaigns and the controls that blunt it. It is written for security teams at South Korean and Japanese government bodies, NGOs, media and universities, and for anyone who protects defectors, activists or researchers working on North Korea. Attribution is as stated by the named sources; the page describes behaviour for defenders only.

Key takeaways

  • People, not networks, are the target. Kaspersky assessed in 2021 that the group targeted individuals rather than organisations, and Zscaler's 2025 victims were individuals tied to North Korea issues, not enterprises.
  • Stolen accounts make the lure credible. Kaspersky found the group messaging a target's acquaintance through the target's stolen Facebook account, then emailing from another stolen account.
  • Local document formats and browser zero-days. HWP (Hangul) documents, Office remote templates and exploits for Flash, Word, Internet Explorer and Edge are all recorded, including CVE-2022-41128.
  • Surveillance goes beyond files. ATT&CK records audio capture and a Bluetooth device harvester; Zscaler reported keystrokes, screenshots every 30 seconds and microphone recording in 2025; Kaspersky found an Android variant collecting SMS, call logs and recordings.
  • Defend the person: phishing-resistant MFA on personal email and social accounts, macro and script restrictions, browser patching and out-of-band checks with contacts.

Who they are (names and aliases)

MITRE ATT&CK (G0067) calls APT37 a North Korean state-sponsored cyber espionage group active since at least 2012, primarily targeting South Korea but also Japan, Vietnam, Russia, Nepal, China, India, Romania, Kuwait and other parts of the Middle East. It links the group to 2016 to 2018 campaigns including Operation Daybreak, Operation Erebus, Golden Time, Evil New Year, FreeMilk and North Korean Human Rights. ATT&CK also warns that North Korean group definitions overlap significantly and that some researchers report all North Korean activity as Lazarus Group.

NameUsed byNotes
APT37 (G0067)Mandiant (FireEye 2018 report); MITRE ATT&CK; GooglePrimary ATT&CK name
ScarCruftKasperskyUsed in Securelist research since 2016
Reaper, TEMP.ReaperMandiantEarlier names in the 2018 report
Group123Cisco TalosListed in ATT&CK
InkySquidVolexityUsed in 2021 browser exploit research
Ricochet ChollimaCrowdStrikeListed in ATT&CK
Ruby SleetNamed by Zscaler (2025) as an aliasNot listed in ATT&CK G0067; confirm against Microsoft's own mapping

Vendor alias tables for North Korean groups are inconsistent; for example, Zscaler's 2025 report also lists Velvet Chollima, a name other sources apply to Kimsuky. Match on indicators. See the Kimsuky Threat Actor Profile, the Lazarus Group Threat Actor Profile and the North Korea Cyber Operations Guide.

Targeting and motivation

  • Defectors and human rights activists: Kaspersky's November 2021 report names defectors, human rights activists, journalists covering North Korea and government bodies tied to the Korean Peninsula.
  • Policymakers and journalists: Google's December 2022 post says APT37 has historically targeted South Korean users, defectors, policymakers, journalists and human rights activists.
  • Academics and researchers: Zscaler's September 2025 report names international affairs, political science, academia and research.
  • Beyond South Korea: ATT&CK lists victims in Japan, Vietnam, Russia, Nepal, China, India, Romania and the Middle East.

The motivation in every source is intelligence collection on people of interest to the North Korean state. That makes personal phones, personal email and social media part of the attack surface.

Tactics and techniques (MITRE ATT&CK)

ATT&CK v20 links APT37 to 29 techniques and to software including ROKRAT, BLUELIGHT, NavRAT, DOGCALL, Final1stspy and Cobalt Strike. All IDs below are listed for G0067, except those marked as mapped from Kaspersky (2021) or Zscaler (2025) reporting.

TacticTechniqueHow it shows up
Initial accessT1566.001 Spearphishing AttachmentMalicious HWP attachments; password-protected RAR archives
Initial accessT1189 Drive-by CompromiseStrategic compromise of South Korean websites; torrent sites
ExecutionT1203 Exploitation for Client ExecutionFlash (CVE-2016-4117, CVE-2018-4878), Word (CVE-2017-0199), Internet Explorer (CVE-2020-1380), Internet Explorer and Edge (CVE-2021-26411, listed by CISA KEV as an Internet Explorer flaw)
ExecutionT1204.002 Malicious File; T1559.002 Dynamic Data Exchange; T1059.005 Visual BasicDocuments and scripts that the victim opens
ExecutionT1059.001 PowerShell; T1218.005 Mshta (Kaspersky 2021, Zscaler 2025)Chinotto PowerShell backdoor; Run key launching mshta to fetch an HTA
PersistenceT1547.001 Registry Run Keys; T1053.005 Scheduled TaskRun keys and tasks for scripts and implants
Privilege escalationT1548.002 Bypass User Account ControlUAC bypass in the first-stage dropper
StealthT1055 Process Injection; T1027.003 Steganography; T1036.001 Invalid Code SignatureROKRAT injected into cmd.exe; shellcode in images; invalid certificate claiming a well-known software company
Credential accessT1555.003 Credentials from Web BrowsersZUMKONG stealer for saved browser passwords
Discovery and collectionT1120 Peripheral Device Discovery; T1123 Audio CaptureBluetooth device harvester; SOUNDWAVE microphone capture
Command and controlT1102.002 Bidirectional CommunicationCloud and social platforms for C2, including Dropbox, Box, pCloud, Yandex and Twitter
ImpactT1561.002 Disk Structure WipeAccess to MBR-wiping malware

How they abuse identities and credentials

  1. Hijack the target's own accounts: Kaspersky describes the actor using a target's stolen Facebook account to message an acquaintance, and stolen credentials to contact and target further people. Phishing-resistant MFA on personal social and email accounts is the most direct defence. See the Phishing-Resistant MFA Guide.
  2. Borrow trust from a known contact: because the message comes from a real friend or colleague, sender checks pass. Verifying unexpected document requests by voice or a second channel breaks the chain.
  3. Harvest browser credentials: ZUMKONG steals saved browser passwords, which on a personal machine often include the same email and social accounts the group then reuses.
  4. Take over the phone: the Android Chinotto variant requested access to contacts, SMS, call logs and recordings. SMS one-time codes are exposed on a compromised phone, so Kaspersky's findings argue against SMS as the second factor for high-risk people. See the Mobile Spyware Infection Response Playbook.
  5. Hide in trusted cloud identities: C2 over Dropbox, Box and similar services uses accounts the attacker registers, so the traffic carries a valid OAuth or API identity that proxies trust.

For the people APT37 targets, there is often no enterprise IT team. The Personal Security for High-Risk Individuals Guide covers practical protections for individuals and the NGOs that support them.

Notable incidents

DateActivitySource
2016Operation Daybreak, Flash zero-day exploitationKaspersky Securelist, cited in ATT&CK
February 2018Mandiant publishes "APT37 (Reaper): The Overlooked North Korean Actor", including use of Flash CVE-2018-4878Cited in ATT&CK G0067
2019Bluetooth device harvester reportedKaspersky Securelist, cited in ATT&CK
2021Browser exploits (including CVE-2021-26411) delivering BLUELIGHTVolexity (as InkySquid), cited in ATT&CK
March to September 2021Surveillance of defectors and activists with Chinotto on Windows and Android, after stolen Facebook and email accounts were used for luresKaspersky, 29 November 2021
October to November 2022Internet Explorer zero-day CVE-2022-41128 exploited through an Itaewon disaster lure; patched 8 November 2022Google Threat Analysis Group, 7 December 2022
From June 2025Rust backdoor Rustonotto, Chinotto and FadeStealer against individuals in South Korea, with keylogging, screenshots and audio captureZscaler ThreatLabz, 8 September 2025

Detection and defence

Controls mapped to behaviours

BehaviourControlNIST SP 800-53 Rev. 5NIST CSF 2.0
Account hijack and contact impersonationPhishing-resistant MFA on work and personal accounts; verify unexpected requests out of bandIA-2, AT-2PR.AA-03, PR.AT-01
Malicious HWP, Office and archive files (T1566.001, T1204.002)Block macros from the internet; open untrusted files in a sandboxed viewer; scan password-protected archives or quarantine themSI-3, SI-8DE.CM-09
Browser and Office exploits (T1203)Fast patching of browsers, Office and document viewers; remove legacy IE engine useSI-2, SA-22PR.PS-02
LNK, CHM, HTA and PowerShell chainsBlock LNK and CHM in mail; restrict mshta.exe; PowerShell Constrained Language Mode and script block loggingCM-7, AU-12PR.PS-01
Browser credential theft (T1555.003)Use a password manager instead of browser storage on high-risk devicesIA-5PR.AA-01
Cloud C2 (T1102.002)Allow file-sharing services only for sanctioned tenants; alert on API use by unusual processesSC-7, SI-4DE.CM-01
Mobile spywareBlock sideloading; review app permissions; use hardened modes on phones of high-risk peopleCM-11PR.PS-05

Phased rollout

  1. Week 1: identify staff, contributors and partners who match APT37 targeting, including defector-support caseworkers and North Korea researchers. Give them a single reporting route.
  2. Weeks 2 to 4: move them to security keys or passkeys for work email, then help them enrol personal email and social media. Remove SMS as a recovery method where possible.
  3. Month 2: block LNK, CHM and HTA attachments, quarantine password-protected archives from external senders, and enable PowerShell script block logging. Pilot application control rules for mshta.exe in audit mode, then enforce, keeping an exception list for any legacy tool that needs it.
  4. Month 3: enforce browser and Office update rings with short deadlines and remove Internet Explorer mode where it is not required.
  5. Ongoing: run awareness that uses current-events lures, since APT37 used a disaster within days of it happening.

What to monitor

  • Office or HWP processes spawning scripts: WINWORD.EXE or Hancom processes launching PowerShell, mshta.exe or cmd.exe.
  • Run keys that launch mshta with a remote URL.
  • Hidden PowerShell with execution policy bypass (Event ID 4104 with script block logging).
  • Microphone access by non-communication applications on Windows.
  • Cloud storage API traffic from processes that are not the official client.

Metrics

  • Percentage of high-risk people with phishing-resistant MFA on work and personal accounts.
  • Percentage of high-risk people without SMS as an account recovery method.
  • Median days to patch browser and Office zero-days on their devices.
  • Number of LNK, CHM and HTA attachments blocked per month.
  • Time from a reported suspicious message to verification with the supposed sender.

Small NGOs and media outlets can cover most of the risk with security keys, a password manager, fast automatic updates and a rule to verify any unexpected file by phone. Government bodies should add application control, centralised script logging and threat hunting, and coordinate with KISA or the national CERT.

Common failure patterns

  • Trusting a message because it came from a friend: Kaspersky's case started from the target's own stolen social account.
  • Using SMS codes on a phone that may be compromised: the Android variant read SMS.
  • Allowing password-protected archives through: the password defeats gateway scanning and adds false credibility.
  • Patching browsers slowly for "low-risk" users: high-risk individuals are exactly who zero-days like CVE-2022-41128 are used against.
  • Focusing only on the corporate network: APT37's victims are often individuals on personal devices.

Practitioner checklist

  • List people who match APT37 targeting and brief them on account-hijack lures.
  • Require phishing-resistant MFA on work accounts and help secure personal accounts.
  • Remove SMS as a second factor or recovery method for high-risk people.
  • Block LNK, CHM and HTA attachments and quarantine encrypted archives from outside.
  • Enable PowerShell script block logging and restrict mshta.exe.
  • Patch browsers, Office and Hangul document viewers on short deadlines.
  • Replace browser password storage with a password manager on high-risk devices.
  • Block sideloaded apps and review phone app permissions.
  • Map APT37, ScarCruft, Reaper, Group123, InkySquid and Ricochet Chollima together in threat intelligence tools.

Standards and references

Related NHI Mgmt Group resources: Kimsuky Threat Actor Profile · North Korea Cyber Operations Guide · Threat Actor Groups Guide · Personal Security for High-Risk Individuals Guide · Mobile Spyware Infection Response Playbook

PDF for subscribers

Take this guide with you

Every section, table and checklist in one PDF, with links to the sources and related guides. Subscribe to our weekly NHI & AI Identity Journal and we’ll email you a download link.

Free weekly newsletter

Subscribe to the NHI & AI Identity Journal

The latest on NHI and Agentic AI security – articles, research, breaches, news and events every week.

Bonus 33% off our NHI Course when you subscribe.

NHIMG Editorial Note
Researched and written by the NHI Mgmt Group editorial team using AI tools, under the editorial direction of Lalit Choda. Facts are drawn from the sources listed on this page and checked by our automated quality process. Last updated 27 October 2026. Spotted an error? Use our contact page.
NHI Mgmt Group, the independent authority on Non-Human Identity, IAM and Agentic AI security. nhimg.org

PDF Downloads for Newsletter Subscribers

Get the guide

Enter your details and we’ll email you a download link.

The full guide is free to read online. The PDF is a thank you for subscribing.

Fill in all fields marked * and tick the PDF box to continue.