← Back to Cybersecurity News Center
Severity
HIGH
CVSS
7.5
Priority
1.000
×
Tip
Pick your view
Analyst for full detail, Executive for the short version.
Analyst
Executive
Executive Summary
CrowdStrike's 2026 Financial Services Threat Landscape Report documents a convergence of nation-state theft, organized cybercrime, and AI-accelerated attacks against global financial institutions at a scale and speed that legacy defenses cannot match. DPRK-affiliated actors stole approximately $2 billion in digital assets during the reporting period, eCrime targeting financial firms surged 27%, and hands-on-keyboard intrusions, where human adversaries operate interactively inside victim environments, increased 43% globally and 48% in North America over two years. The report signals that financial sector adversaries have matured from opportunistic, tool-driven campaigns to persistent, human-directed operations that exploit trust relationships, AI-generated lures, and supply chain access vectors simultaneously.
Impact Assessment
CISA KEV Status
Not listed
Threat Severity
HIGH
High severity — prioritize for investigation
Actor Attribution
HIGH
MURKY PANDA, Lazarus Group (DPRK-nexus)
TTP Sophistication
HIGH
17 MITRE ATT&CK techniques identified
Detection Difficulty
HIGH
Multiple evasion techniques observed
Target Scope
INFO
Financial services organizations broadly; cryptocurrency exchanges; fintech platforms; Microsoft 365 environments (MURKY PANDA targeting via trusted-relationship cloud intrusions); no specific CVE-linked products identified
Are You Exposed?
⚠
Your industry is targeted by MURKY PANDA, Lazarus Group (DPRK-nexus) → Heightened risk
⚠
You use products/services from Financial services organizations broadly; cryptocurrency exchanges; fintech platforms; Microsoft 365 environments (MURKY PANDA targeting via trusted-relationship cloud intrusions); no specific CVE-linked products identified → Assess exposure
⚠
17 attack techniques identified — review your detection coverage for these TTPs
✓
Your EDR/XDR detects the listed IOCs and TTPs → Reduced risk
✓
You have incident response procedures for this threat type → Prepared
Assessment estimated from severity rating and threat indicators
Business Context
The financial sector threat environment described in this report creates direct exposure across revenue continuity, regulatory standing, and institutional trust. Ransomware targeting financial institutions — up 27% in eCrime activity — carries immediate operational impact: system downtime, transaction processing disruption, and potential notification obligations that trigger regulatory scrutiny under financial sector data protection frameworks. DPRK cryptocurrency theft at $2 billion scale and MURKY PANDA's espionage-focused cloud intrusions signal that financial institutions face adversaries with nation-state resources and long-term persistence objectives, not simply opportunistic attackers seeking quick monetization. For boards, the practical implication is that cyber risk is no longer bounded by the strength of the organization's own controls — supply chain and trusted vendor relationships now represent the primary unmanaged attack surface in the financial sector.
You Are Affected If
Your organization operates in financial services, fintech, cryptocurrency exchange, or digital asset custody
Your Microsoft 365 environment is accessible to third-party vendors, managed service providers, or supplier accounts
Your organization uses or integrates with cryptocurrency platforms, DeFi protocols, or digital asset infrastructure with software supply chain dependencies
Your security operations rely primarily on automated or signature-based detection without behavioral analytics for interactive intrusion activity
Your organization has not formally assessed trusted-relationship cloud access paths as an attack vector in your current threat model
Board Talking Points
Nation-state actors stole $2 billion from financial sector targets during the past year, and organized cybercrime against financial institutions surged 27% — this sector is the primary target, not a collateral one.
Within 30 days, commission a review of all third-party vendor access to your Microsoft 365 and core financial systems, and verify that behavioral detection capabilities — not just signature-based tools — are operational against interactive intrusion techniques.
Organizations that do not address supply chain and trusted-relationship access vectors now face the same attack surface that enabled nation-state espionage and multi-week ransomware dwell operations documented in this report, with direct exposure to regulatory enforcement and reputational damage following a breach.
DORA (EU Digital Operational Resilience Act) — financial entities operating in the EU face direct ICT risk management, incident reporting, and third-party provider oversight obligations triggered by the supply chain intrusion and ransomware patterns documented in this report
FFIEC Cybersecurity Framework — U.S. financial institutions are subject to FFIEC guidance on authentication, access management, and incident response; the trusted-relationship cloud intrusions and session token theft TTPs documented here directly implicate FFIEC authentication and vendor management expectations
SEC Cybersecurity Disclosure Rules — publicly traded financial institutions subject to SEC rules face material incident disclosure obligations if the ransomware or espionage activity described reaches the materiality threshold; the 27% eCrime surge and interactive intrusion increase are relevant to materiality risk assessments
Technical Analysis
CrowdStrike's April 2025 to March 2026 reporting period reveals a financial sector threat environment defined by three converging forces: state-sponsored cryptocurrency theft at industrial scale, commoditized eCrime infrastructure expanding rapidly, and nation-state espionage pivoting to cloud trust chains.
The most operationally significant shift is the 43% global and 48% North American increase in hands-on-keyboard intrusions against financial institutions over two years.
This metric matters because interactive intrusions, where a human operator navigates an environment in real time, reduce detection and response time between initial access and payload execution, evade automated detection tuned for scripted behavior, and enable adversaries to adapt mid-operation when defenses respond.
Signature-based and rule-heavy detection stacks are structurally disadvantaged against this attack pattern.
DPRK-nexus actors, including clusters affiliated with Lazarus Group, continued their multi-year campaign to generate hard currency through cryptocurrency theft, accounting for approximately $2 billion in stolen digital assets during the period. Their tradecraft consistently involves spear-phishing (T1566 , T1566.002 ), credential theft targeting session cookies (T1539 ), supply chain compromise (T1195 , T1195.002 ), and use of proxy infrastructure (T1090.002 ) to obscure origin. CWE-494 (download of code without integrity check) and CWE-829 (inclusion of functionality from untrusted control sphere) are structural vulnerabilities these actors exploit in cryptocurrency platform software and DeFi protocols.
China-nexus adversary MURKY PANDA represents a distinct but equally significant threat vector. The group conducted trusted-relationship cloud intrusions targeting Microsoft 365 environments, exploiting third-party and supplier trust chains (T1199 , T1195 ) for espionage-focused persistent access. Their approach leverages valid accounts (T1078 ) obtained through compromised vendors, making detection dependent on behavioral analytics rather than signature matching. CWE-426 (untrusted search path) is consistent with their observed DLL hijacking technique (T1574.001 ). The use of trusted relationships as an access vector means organizations with strong perimeter controls remain exposed through their supply chain.
AI's role in this threat landscape is accelerating adversary capabilities across multiple phases: generating convincing spear-phishing content at scale (T1566 , T1598 ), automating reconnaissance (T1590 ), and accelerating custom malware development (T1587.001 , T1588 ). AI-generated social engineering removes the grammatical and contextual errors that historically flagged phishing attempts, lowering the detection rate of human-targeted lures against financial sector employees.
The convergence of these three vectors, DPRK financial theft, MURKY PANDA espionage, and AI-amplified eCrime, against the same sector creates compounding risk. An organization compromised through a trusted vendor relationship (MURKY PANDA) may simultaneously face ransomware operators (T1486 ) who acquired access through the same initial foothold via access broker markets (T1650 ).
Action Checklist IR ENRICHED
Triage Priority:
URGENT
Escalate immediately to CISO, legal counsel, and FS-ISAC if Entra ID audit logs confirm unauthorized OAuth consent grants to external applications, session token replay events are detected from DPRK-geolocated IPs, any digital asset custody system integrity check fails, or if ransomware staging artifacts (bulk archive creation, shadow copy deletion via vssadmin) are detected on financial systems — all of which trigger potential SAR filing obligations under FinCEN guidance and breach notification timelines under applicable state and federal financial regulations.
1
Step 1: Assess supply chain exposure, inventory all third-party vendors, managed service providers, and SaaS platforms with access to your Microsoft 365 environment or financial systems; cross-reference against trusted-relationship intrusion TTPs (T1199, T1195) observed in MURKY PANDA activity; apply NIST AC-20 (Use of External Systems) to enforce documented terms and access controls for each external system connection
IR Detail
Preparation
NIST 800-61r3 §2 — Preparation: establishing IR capability and hardening posture before MURKY PANDA-style trusted-relationship intrusions materialize
NIST AC-20 (Use of External Systems)
NIST AC-3 (Access Enforcement)
NIST AC-17 (Remote Access)
NIST RA-3 (Risk Assessment) — implied for third-party risk scoring
CIS 1.1 (Establish and Maintain Detailed Enterprise Asset Inventory)
CIS 2.1 (Establish and Maintain a Software Inventory)
Compensating Control
Export the Microsoft 365 Entra ID (Azure AD) connected applications list via PowerShell: `Get-MgApplication | Select-Object DisplayName, AppId, SignInAudience` and cross-reference each entry against your vendor register. For MSP OAuth token grants specifically, run `Get-MgOauth2PermissionGrant | Select-Object ClientId, Scope, ConsentType` to surface delegated permissions MURKY PANDA abuses via trusted-partner OAuth paths. Flag any 'AllPrincipals' consent grants for immediate review. Maintain findings in a shared spreadsheet reviewed weekly by both team members.
Preserve Evidence
Before restricting any third-party connection, capture: (1) Microsoft 365 Unified Audit Log entries under 'AzureActiveDirectory' workload — specifically 'Add delegation entry' and 'Consent to application' operations that predate your review window; (2) Entra ID Sign-In Logs filtered for service principal sign-ins from unexpected IP geolocation or ASNs inconsistent with the vendor's known infrastructure; (3) Microsoft 365 admin audit log for 'Set-PartnerInformation' or 'New-PartnerAccess' events indicating delegated admin privilege grants consistent with MURKY PANDA trusted-relationship initial access via T1199.
2
Step 2: Audit interactive intrusion detection capability, evaluate whether your SIEM and EDR stack can detect hands-on-keyboard behavior (T1059, T1560, T1090.002) versus only automated attack tooling; review AU-6 (Audit Record Review, Analysis, and Reporting) compliance and ensure analysts are reviewing logs at a frequency sufficient to catch dwell-time intrusions; CIS 8.2 (Collect Audit Logs) should be verified across all financial system endpoints and cloud workloads
IR Detail
Detection & Analysis
NIST 800-61r3 §3.2 — Detection and Analysis: validating that monitoring capability can surface human-operated intrusion activity characteristic of DPRK-affiliated and eCrime hands-on-keyboard operators
NIST AU-2 (Event Logging)
NIST AU-6 (Audit Record Review, Analysis, and Reporting)
NIST AU-12 (Audit Record Generation)
NIST SI-4 (System Monitoring) — implied for behavioral detection
CIS 8.2 (Collect Audit Logs)
Compensating Control
Deploy Sysmon with SwiftOnSecurity's configuration (https://github.com/SwiftOnSecurity/sysmon-config) on all financial system endpoints; it generates Event ID 1 (Process Create) and Event ID 3 (Network Connect) entries that expose T1059 shell invocations and T1090.002 proxy tool execution. Write Sigma rules targeting cmd.exe or powershell.exe spawned by financial application parent processes (e.g., java.exe, w3wp.exe, svchost.exe hosting line-of-business apps). Use `Get-WinEvent -LogName Security -FilterXPath` queries to pull Event ID 4688 (Process Creation) with `/CommandLine` auditing enabled via Group Policy (Audit Process Creation → include command line). For T1560 archive staging, alert on `7z.exe`, `rar.exe`, or `Compress-Archive` invocations writing to unusual output paths.
Preserve Evidence
Before tuning detection rules, preserve: (1) Windows Security Event Log — Event ID 4688 (Process Creation with command line) showing interactive shell chains inconsistent with financial application behavior; (2) Sysmon Event ID 1 logs capturing parent-child process trees where a legitimate financial platform process spawns cmd.exe, wscript.exe, or mshta.exe — the hallmark of hands-on-keyboard T1059 execution used in the intrusion pattern documented in the CrowdStrike report; (3) Microsoft 365 Unified Audit Log 'SearchQueryInitiated' and 'FileDownloaded' events indicating bulk data staging consistent with T1560 pre-exfiltration behavior by eCrime operators; (4) NetFlow or Windows Firewall logs showing outbound connections on non-standard ports from financial workstations to IPs consistent with T1090.002 proxy infrastructure.
3
Step 3: Harden credential and identity controls against session hijacking and valid account abuse, enforce MFA on all externally exposed applications and administrative accounts per CIS 6.3, 6.4, and 6.5; implement and enforce MFA and credential rotation practices to limit the value of stolen session tokens (T1539) and valid credentials (T1078); review AC-2 (Account Management) to ensure dormant and third-party accounts are disabled per CIS 5.3
IR Detail
Containment
NIST 800-61r3 §3.3 — Containment: limiting adversary persistence and lateral movement by invalidating DPRK and eCrime actors' stolen session tokens (T1539) and credential reuse (T1078) before full eradication is confirmed
NIST AC-2 (Account Management)
NIST AC-7 (Unsuccessful Logon Attempts)
NIST AC-12 (Session Termination)
NIST IA-5 (Authenticator Management) — implied for credential rotation
CIS 5.3 (Disable Dormant Accounts)
CIS 6.3 (Require MFA for Externally-Exposed Applications)
CIS 6.4 (Require MFA for Remote Network Access)
CIS 6.5 (Require MFA for Administrative Access)
Compensating Control
For teams without a PAM solution: (1) Use Microsoft Entra ID's built-in Conditional Access (free tier via Security Defaults) to enforce MFA on all admin accounts immediately; (2) Run `Get-MgUser -Filter 'accountEnabled eq true' | Where-Object {$_.LastSignInDateTime -lt (Get-Date).AddDays(-45)}` to identify dormant accounts for CIS 5.3 enforcement; (3) Force token revocation for all accounts with M365 access via `Revoke-MgUserSignInSession -UserId <UPN>` — critical for invalidating stolen session tokens exploited via T1539 without requiring full password reset; (4) Audit service accounts with non-expiring passwords via `Get-MgUser -Filter 'passwordPolicies eq DisablePasswordExpiration'` and rotate credentials immediately for any with M365 or financial system access.
Preserve Evidence
Before rotating credentials, preserve: (1) Microsoft Entra ID Sign-In Logs — filter for 'Interrupted' or 'Success' sign-ins using legacy authentication protocols (Basic Auth, IMAP, POP3) which bypass MFA and are the primary vector for T1078 valid account abuse in M365 environments; (2) Microsoft 365 Unified Audit Log entries for 'UserLoggedIn' events from unfamiliar User-Agent strings or impossible-travel IP pairs indicating session token replay (T1539); (3) Entra ID audit log for 'Update user' or 'Reset user password' events initiated by non-admin accounts, indicating adversary credential manipulation; (4) Browser artifact dumps (Chrome/Edge `Cookies`, `Login Data`, `Web Data` SQLite files at `%LOCALAPPDATA%\Google\Chrome\User Data\Default\`) from potentially compromised financial workstations, as DPRK-affiliated actors have demonstrated infostealer use to harvest session cookies prior to T1539 replay.
4
Step 4: Integrate cryptocurrency and digital asset exposure into your threat model, if your organization operates, custodies, or interfaces with digital asset platforms, map DPRK-affiliated TTPs (T1566.002, T1195.002, T1539, CWE-494, CWE-829) into your threat register; assess whether software supply chain integrity checks are enforced for any wallet, exchange, or DeFi integration code
IR Detail
Preparation
NIST 800-61r3 §2 — Preparation: proactively mapping DPRK-affiliated TTPs against digital asset infrastructure before the $2B theft pattern documented in the CrowdStrike 2026 report is replicated against your organization
NIST RA-3 (Risk Assessment) — implied for threat model integration
NIST SA-12 (Supply Chain Protection) — implied for software supply chain integrity
NIST SI-7 (Software, Firmware, and Information Integrity)
NIST AC-20 (Use of External Systems)
CIS 2.2 (Ensure Authorized Software is Currently Supported)
CIS 7.1 (Establish and Maintain a Vulnerability Management Process)
Compensating Control
Use OSSEC or Wazuh (free SIEM/FIM) to implement file integrity monitoring on any wallet client binaries, exchange SDK libraries, or DeFi smart contract deployment scripts — flagging unauthorized modifications consistent with CWE-494 (Download of Code Without Integrity Check) exploitation. For supply chain integrity: enforce SHA-256 hash verification of third-party wallet SDK releases against vendor-published checksums before deployment. Write a YARA rule targeting known DPRK-associated packer signatures (e.g., Lazarus Group's custom packers documented in CISA advisories AA21-048A and AA22-108A) to scan any newly integrated cryptocurrency library binaries. For T1566.002 (spear-phishing link to crypto platform impersonation), configure DNS RPZ or Pi-hole blocklists using CISA and FS-ISAC published DPRK phishing domain feeds.
Preserve Evidence
Before updating the threat register, preserve: (1) Git commit history and CI/CD pipeline logs for any wallet integration, exchange API client, or DeFi contract code — DPRK supply chain compromises via T1195.002 leave evidence of unauthorized commits or modified build artifacts inconsistent with developer activity patterns; (2) NPM or PyPI package installation logs (npm audit log, pip install history from `~/.local/lib` timestamps) for cryptocurrency-related dependencies, as DPRK actors have poisoned package repositories used by crypto exchanges; (3) Network traffic captures (PCAP) of outbound connections from wallet or exchange integration servers — DPRK C2 infrastructure communicates over HTTPS to domains mimicking legitimate CDN or cloud providers; (4) Binary hash values of all currently deployed wallet client and exchange SDK executables for baseline comparison against future tampering consistent with CWE-829 (Inclusion of Functionality from Untrusted Control Sphere).
5
Step 5: Evaluate AI-generated phishing resilience, test your organization's detection and employee response capability against AI-generated spear-phishing (T1566, T1598) that lacks traditional grammatical indicators; verify that email security controls are tuned for behavioral and contextual signals, not only known-bad signatures; brief leadership on the specific financial sector eCrime surge (27%) and ransomware victim count increase documented in this report with organizational risk context, not generic threat landscape language
IR Detail
Preparation
NIST 800-61r3 §2 — Preparation: building organizational and technical resilience against AI-augmented phishing campaigns driving the 27% eCrime surge against financial sector targets documented in the CrowdStrike 2026 report
NIST SI-8 (Spam Protection) — implied for email filtering controls
NIST AT-2 (Literacy Training and Awareness) — implied for employee phishing resilience
NIST AU-6 (Audit Record Review, Analysis, and Reporting)
CIS 7.1 (Establish and Maintain a Vulnerability Management Process)
CIS 8.2 (Collect Audit Logs)
Compensating Control
Run free phishing simulation using GoPhish (https://getgophish.com) with AI-polished lure templates that mimic wire transfer authorization requests, crypto custody notifications, or M365 OAuth consent prompts — the three highest-volume financial sector lure types in this threat period. For email gateway controls without a commercial SEG: enable Microsoft Defender for Office 365's free-tier 'Safe Links' and configure DMARC/DKIM/SPF enforcement via your DNS provider and MXToolbox validation. Extract Microsoft 365 email headers for suspicious messages and parse DKIM alignment, SPF result, and sender IP geolocation using the free MessageHeader analyzer at MXToolbox. Write Sigma rules against Microsoft 365 Unified Audit Log 'MailItemsAccessed' and 'Send' operations to detect post-phishing account takeover activity.
Preserve Evidence
Before tuning email controls, preserve: (1) Microsoft 365 Message Trace logs and email header data for any suspected AI-generated phishing messages targeting financial staff — specifically capture 'X-Originating-IP', DKIM signature alignment, and Return-Path mismatches that remain even in grammatically perfect AI-generated lures; (2) Microsoft 365 Unified Audit Log 'MailItemsAccessed' events (E3/E5 audit) showing inbox access from unfamiliar IP addresses or User-Agent strings within minutes of a phishing link click, indicating rapid account takeover following T1566 credential harvesting; (3) Microsoft Entra ID 'Risky Sign-In' log entries correlated temporally with phishing simulation or real-incident click events; (4) Browser process tree via Sysmon Event ID 1 showing browser.exe spawning unexpected child processes (e.g., powershell.exe, wscript.exe) following a phishing link click — the execution chain for T1598 credential harvesting pages that deploy additional malware.
6
Step 6: Monitor CrowdStrike's published IOC feeds and adversary profile updates, track CrowdStrike Intelligence releases for MURKY PANDA and DPRK-affiliated cluster indicators; subscribe to CISA financial sector alerts and FS-ISAC advisories for downstream regulatory guidance or law enforcement actions related to this report period's findings
IR Detail
Post-Incident
NIST 800-61r3 §4 — Post-Incident Activity: sustaining threat intelligence integration to detect recurrence of MURKY PANDA and DPRK-affiliated activity and incorporate lessons from the CrowdStrike 2026 reporting period into updated detection and response posture
NIST IR-4 (Incident Handling)
NIST SI-5 (Security Alerts, Advisories, and Directives)
NIST AU-6 (Audit Record Review, Analysis, and Reporting)
NIST AU-13 (Monitoring for Information Disclosure)
CIS 7.1 (Establish and Maintain a Vulnerability Management Process)
CIS 7.2 (Establish and Maintain a Remediation Process)
Compensating Control
Operationalize free threat intelligence feeds using MISP (Malware Information Sharing Platform, open-source) or OpenCTI to ingest CISA Known Exploited Vulnerabilities catalog, FS-ISAC TLP:WHITE indicator feeds, and CrowdStrike's public Adversary Universe blog posts. Configure automated correlation of incoming DPRK-affiliated IOCs (IP, domain, file hash) against Microsoft 365 Unified Audit Logs and Sysmon logs using osquery scheduled queries or a free Elastic SIEM instance. Set a Google Alert or RSS feed subscription for CISA AA advisories tagged 'North Korea' and 'financial services' to receive regulatory-grade threat intelligence at no cost. Maintain a running MURKY PANDA TTP matrix in ATT&CK Navigator (free, browser-based) updated each time CrowdStrike or CISA publishes new activity cluster details.
Preserve Evidence
For ongoing post-incident monitoring, preserve and continuously collect: (1) Microsoft 365 Unified Audit Log exports retained for minimum 90 days (365 days if E5 licensed) — the primary forensic record for detecting MURKY PANDA trusted-relationship re-entry via T1199 after initial remediation; (2) DPRK-affiliated C2 domain and IP indicators from CISA advisories cross-referenced against DNS query logs (Windows DNS Debug Log or Sysmon Event ID 22 — DNS Query) from financial system endpoints; (3) FS-ISAC shared indicator reports for DPRK cryptocurrency theft campaigns — match file hashes against endpoint AV/EDR quarantine logs and Sysmon Event ID 7 (Image Load) records; (4) Entra ID audit log entries for any re-establishment of OAuth delegated grants or partner access following the remediation actions in Steps 1–3, indicating adversary re-entry attempt via the same trusted-relationship vector documented in MURKY PANDA activity.
Recovery Guidance
Following credential rotation and third-party access remediation, continuously monitor Microsoft 365 Unified Audit Log for 'Consent to application' and service principal sign-in events for a minimum of 30 days to detect MURKY PANDA re-entry via new OAuth paths or re-compromised MSP partner accounts. Verify supply chain integrity of all cryptocurrency-adjacent libraries and wallet SDKs by re-hashing deployed binaries against vendor-published checksums before restoring any digital asset platform connectivity. Sustain elevated logging fidelity (Sysmon, M365 full audit) and weekly analyst review of hands-on-keyboard behavioral indicators for at least 60 days post-containment, as DPRK-affiliated actors have demonstrated extended dwell times and deliberate re-entry following incomplete remediation.
Key Forensic Artifacts
Microsoft 365 Unified Audit Log — 'Consent to application', 'Add delegation entry', and 'UserLoggedIn' events with legacy authentication User-Agent strings, evidencing MURKY PANDA trusted-relationship intrusion via T1199 and T1078 valid account abuse against M365 tenants
Microsoft Entra ID Sign-In Logs — service principal authentication events from IP addresses or ASNs inconsistent with registered MSP or SaaS vendor infrastructure, surfacing DPRK-affiliated actor use of compromised partner credentials for cloud lateral movement
Sysmon Event ID 1 (Process Create) logs — parent-child process chains where financial application processes (java.exe, w3wp.exe, node.exe for DeFi integrations) spawn cmd.exe, powershell.exe, or wscript.exe, the hallmark execution artifact of hands-on-keyboard T1059 activity documented in the CrowdStrike 2026 intrusion pattern analysis
Browser session cookie stores and credential databases (Chrome/Edge SQLite files at %LOCALAPPDATA%\Google\Chrome\User Data\Default\Cookies and Login Data) on financial workstations — primary artifact of DPRK infostealer pre-positioning for T1539 session token theft enabling subsequent cryptocurrency platform account takeover
Git repository commit history and CI/CD pipeline build logs for cryptocurrency wallet, exchange SDK, or DeFi smart contract codebases — forensic record of unauthorized commits or modified build artifacts consistent with DPRK supply chain compromise via T1195.002 and CWE-494 exploitation targeting the $2B digital asset theft campaign documented in this report
Detection Guidance
Detection for the TTPs documented in this report requires behavioral analytics prioritized over signature matching, particularly given the 43% increase in interactive intrusions.
For MURKY PANDA trusted-relationship cloud intrusions (T1199 , T1078 , T1574.001 ): Hunt for authentication events in Microsoft 365 audit logs originating from vendor or supplier accounts at unusual hours or from unexpected geographies.
Flag service accounts authenticating interactively.
Monitor for DLL hijacking indicators, unexpected DLL load paths for legitimate applications, particularly in directories writable by non-administrative users (D3-SFA: System File Analysis). Review Azure AD sign-in logs for token reuse anomalies consistent with session cookie theft (T1539 ). CIS 8.2 compliance across cloud workloads is a prerequisite for this hunt.
For DPRK cryptocurrency theft TTPs (T1566.002 , T1195 , T1195.002 , CWE-494): Monitor software build and deployment pipelines for unsigned or unverified code introduced through dependency updates, CWE-494 and CWE-829 patterns appear in supply chain compromise of DeFi and exchange platforms. Hunt for anomalous outbound connections (T1090.002 ) from systems handling digital asset transactions. Flag large or unusual asset transfer events against behavioral baselines. D3-FMBV (File Magic Byte Verification) can detect payload masquerading in download chains.
For AI-amplified phishing and social engineering (T1566 , T1598 , T1590 ): Shift email security tuning toward sender behavioral signals, domain age, sending infrastructure reputation, and contextual anomalies in request content, rather than grammatical error detection, which AI-generated lures no longer trigger. Monitor for credential harvesting infrastructure (T1598 ) targeting financial sector employees via lookalike domains.
For hands-on-keyboard intrusion indicators (T1059 , T1560 , T1486 ): Alert on interactive command execution from service or non-interactive accounts. Monitor for staging and archiving behaviors (T1560 ), file compression utilities executing in unusual contexts, particularly on systems holding financial data. Ransomware deployment (T1486 ) in financial environments is increasingly preceded by multi-week dwell periods; look for lateral movement and privilege escalation in the weeks before encryption events. D3-LAM (Local Account Monitoring) and D3-UAP (User Account Permissions) reviews support this detection posture.
Log sources to prioritize: Microsoft 365 Unified Audit Log, Azure AD sign-in and provisioning logs, EDR telemetry for process creation and DLL load events, network proxy logs for external connections from financial transaction systems, and email gateway logs with full header visibility.
Indicators of Compromise (1)
Export as
Splunk SPL
KQL
Elastic
Copy All (1)
1 tool
Type Value Enrichment Context Conf.
⚙ TOOL
Pending — refer to CrowdStrike 2026 Financial Services Threat Landscape Report for published indicators
CrowdStrike's report documents MURKY PANDA and DPRK-affiliated cluster TTPs including C2 infrastructure, payload hashes, and DLL hijacking artifacts; specific indicator values are not reproduced in the public blog summary and should be retrieved directly from the full report or CrowdStrike Intelligence portal
LOW
Platform Playbooks
Microsoft Sentinel / Defender
CrowdStrike Falcon
AWS Security
🔒
Microsoft 365 E3
3 log sources
Basic identity + audit. No endpoint advanced hunting. Defender for Endpoint requires separate P1/P2 license.
🛡
Microsoft 365 E5
18 log sources
Full Defender suite: Endpoint P2, Identity, Office 365 P2, Cloud App Security. Advanced hunting across all workloads.
🔍
E5 + Sentinel
27 log sources
All E5 tables + SIEM data (CEF, Syslog, Windows Security Events, Threat Intelligence). Analytics rules, playbooks, workbooks.
Hard indicator (direct match)
Contextual (behavioral query)
Shared platform (review required)
IOC Detection Queries (1)
Known attack tool — NOT a legitimate system binary. Any execution is suspicious.
KQL Query Preview
Read-only — detection query only
// Threat: DPRK Steals $2B in Crypto, eCrime Surges 27%, and AI Amplifies Financial Sector
// Attack tool: Pending — refer to CrowdStrike 2026 Financial Services Threat Landscape Report for published indicators
// Context: CrowdStrike's report documents MURKY PANDA and DPRK-affiliated cluster TTPs including C2 infrastructure, payload hashes, and DLL hijacking artifacts; specific indicator values are not reproduced in th
DeviceProcessEvents
| where Timestamp > ago(30d)
| where FileName =~ "Pending — refer to CrowdStrike 2026 Financial Services Threat Landscape Report for published indicators"
or ProcessCommandLine has "Pending — refer to CrowdStrike 2026 Financial Services Threat Landscape Report for published indicators"
or InitiatingProcessCommandLine has "Pending — refer to CrowdStrike 2026 Financial Services Threat Landscape Report for published indicators"
| project Timestamp, DeviceName, FileName, FolderPath,
ProcessCommandLine, AccountName, AccountDomain,
InitiatingProcessFileName, InitiatingProcessCommandLine
| sort by Timestamp desc
MITRE ATT&CK Hunting Queries (5)
Sentinel rule: Phishing email delivery
KQL Query Preview
Read-only — detection query only
EmailEvents
| where Timestamp > ago(7d)
| where ThreatTypes has "Phish" or DetectionMethods has "Phish"
| summarize Attachments = make_set(AttachmentCount), Urls = make_set(UrlCount) by NetworkMessageId, Timestamp, SenderFromAddress, RecipientEmailAddress, Subject, DeliveryAction, DeliveryLocation, ThreatTypes
| sort by Timestamp desc
Sentinel rule: Suspicious PowerShell command line
KQL Query Preview
Read-only — detection query only
DeviceProcessEvents
| where Timestamp > ago(7d)
| where FileName in~ ("powershell.exe", "pwsh.exe", "cmd.exe", "wscript.exe", "cscript.exe", "mshta.exe")
| where ProcessCommandLine has_any ("-enc", "-nop", "bypass", "hidden", "downloadstring", "invoke-expression", "iex", "frombase64", "new-object net.webclient")
| project Timestamp, DeviceName, FileName, ProcessCommandLine, AccountName, InitiatingProcessFileName
| sort by Timestamp desc
Sentinel rule: Sign-ins from unusual locations
KQL Query Preview
Read-only — detection query only
SigninLogs
| where TimeGenerated > ago(7d)
| where ResultType == 0
| summarize Locations = make_set(Location), LoginCount = count(), DistinctIPs = dcount(IPAddress) by UserPrincipalName
| where array_length(Locations) > 3 or DistinctIPs > 5
| sort by DistinctIPs desc
Sentinel rule: Ransomware activity
KQL Query Preview
Read-only — detection query only
DeviceFileEvents
| where Timestamp > ago(7d)
| where ActionType == "FileRenamed"
| where FileName endswith_any (".encrypted", ".locked", ".crypto", ".crypt", ".enc", ".ransom")
| summarize RenamedFiles = count() by DeviceName, InitiatingProcessFileName, bin(Timestamp, 5m)
| where RenamedFiles > 20
| sort by RenamedFiles desc
Sentinel rule: Supply chain / cross-tenant access
KQL Query Preview
Read-only — detection query only
SigninLogs
| where TimeGenerated > ago(7d)
| where HomeTenantId != ResourceTenantId
| project TimeGenerated, UserPrincipalName, AppDisplayName, IPAddress, Location, HomeTenantId, ResourceTenantId
| sort by TimeGenerated desc
No actionable IOCs for CrowdStrike import (benign/contextual indicators excluded).
No hard IOCs available for AWS detection queries (contextual/benign indicators excluded).
Compliance Framework Mappings
T1566.002
T1539
T1566
T1587.001
T1195
T1590
+11
AT-2
SC-7
SI-3
SI-4
SI-8
CA-7
+13
RS.MI-01
GV.SC-01
DE.CM-01
164.308(a)(7)(ii)(A)
164.308(a)(5)(i)
A.5.29
A.5.34
A.5.21
A.5.23
MITRE ATT&CK Mapping
T1539
Steal Web Session Cookie
credential-access
T1566
Phishing
initial-access
T1195
Supply Chain Compromise
initial-access
T1590
Gather Victim Network Information
reconnaissance
T1560
Archive Collected Data
collection
T1059
Command and Scripting Interpreter
execution
T1650
Acquire Access
resource-development
T1598
Phishing for Information
reconnaissance
T1078
Valid Accounts
defense-evasion
T1195.002
Compromise Software Supply Chain
initial-access
T1588
Obtain Capabilities
resource-development
T1486
Data Encrypted for Impact
impact
T1199
Trusted Relationship
initial-access
Guidance Disclaimer
The analysis, framework mappings, and incident response recommendations in this intelligence
item are derived from established industry standards including NIST SP 800-61, NIST SP 800-53,
CIS Controls v8, MITRE ATT&CK, and other recognized frameworks. This content is provided
as supplemental intelligence guidance only and does not constitute professional incident response
services. Organizations should adapt all recommendations to their specific environment, risk
tolerance, and regulatory requirements. This material is not a substitute for your organization's
official incident response plan, legal counsel, or qualified security practitioners.
View All Intelligence →