Threat reportAPTTL-2026-4755

North Korean Threat Actors Operationalize AI Across Tradecraft (Jade/Coral/Jasper/Diamond/Onyx/Pearl/Sapphire/Emerald Sleet; ProcmacSteal)

highACTIVE

North Korean Threat Actors Operationalize AI Across (TL-2026-4755), also tracked as AI-enabled North Korean tradecraft, is a high-severity advanced persistent threat campaign, first published 2026-10-29. It is attributed to Jade Sleet (North Korea) with high confidence, affects Rust / crates.io arrayref, maps to 13 MITRE ATT&CK techniques (T1027, T1027.004, T1059.001), and is covered by 9 detection rules and 28 indicators of compromise.

Severity
HIGHAssessed severity
CVEs
0None referenced
Techniques
13MITRE ATT&CK
Actors
4Jade Sleet
Detection rules
9SPL · KQL · Sigma
IOCs
28Indicators of compromise

Key facts for TL-2026-4755

Threat ID
TL-2026-4755
Also known as
AI-enabled North Korean tradecraft, ProcmacSteal
Severity
HIGH
Status
ACTIVE
Category
APT
First published
Last reviewed
Attribution
Jade Sleet, Coral Sleet, Jasper Sleet, Lazarus Group
Attribution confidence
HIGH
Nation-state nexus
North Korea
Motivation
FINANCIAL
Target sectors
finance, cryptocurrency, technology, software-supply-chain, government administration, education, energy, telecoms, defense
Target regions
Global
Detection rules
9
Indicators of compromise
28

Malware and tooling in North Korean Threat Actors Operationalize AI Across

Malware and tooling: Diamond, Trojan:Cargo/ProcmacSteal, Trojan:Linux/ProcmacSteal, Trojan:MacOS/ProcmacSteal, Trojan:PowerShell/ProcmacSteal, ARTEX, Cursor

How North Korean Threat Actors Operationalize AI Across works

Microsoft reports eight North Korean threat actors using AI for tailored phishing and recruiter lures, fake personas and resumes, target research, malware development and debugging, synthetic voice/video media, and trojanizing private code repositories. The report names the ProcmacSteal malware family (Linux, PowerShell, MacOS, Cargo variants) tied to Sapphire Sleet's August 2026 Rust software supply chain compromise.

Microsoft Threat Intelligence (2026-10-29) describes a staged adoption of AI by North Korean threat actors. Jade Sleet, Coral Sleet and Jasper Sleet are described as having operationalized AI; Diamond Sleet, Onyx Sleet, Pearl Sleet and Sapphire Sleet as accelerating; and Emerald Sleet as expanding. Motivations span revenue generation for the regime, strategic intelligence collection, and politically motivated disruptive or destructive objectives. Targets include cryptocurrency and financial organizations, IT companies and software supply chains, government, education, oil and gas, and communications/internet policy organizations.

AI is used across the attack lifecycle. For reconnaissance, actors research organizations and employees to craft bespoke lures, and research wallets, exchanges, DeFi, vulnerabilities, APIs and cloud services. For social engineering and identity deception, actors generate tailored phishing and recruiter lures, resumes, biographies and interview answers, enhance or alter profile images, manage multi-platform personas, and (Emerald Sleet) produce synthetic voice and video. Between 2026-06-08 and 2026-06-12 Jade Sleet ran job-themed targeting through three front companies with tailored domains. In August-September 2026 Onyx Sleet generated over 100 target-specific PDF lure files, producing dozens within minutes.

For malware and tooling, actors use AI for code generation, debugging, refactoring, script, command and persistence development, obfuscation and cross-language adaptation. Coral Sleet has operated a GitHub repository trojanization workflow since April 2026. Sapphire Sleet's August 2026 Rust compromise incorporated dynamic .NET compilation for defense evasion and is associated with the Trojan:Linux/ProcmacSteal, Trojan:PowerShell/ProcmacSteal, Trojan:MacOS/ProcmacSteal and Trojan:Cargo/ProcmacSteal detections. Third-party analysis (StepSecurity, Semgrep) of the 2026-08-20 crates.io incident shows compromised arrayref 0.3.10, internment 0.8.7 and append-only-vec 0.1.9 pulling the typosquatted proc-macro1 crate, whose build.rs downloaded and detached a platform-specific payload from Hostwinds-hosted infrastructure during cargo build; Semgrep and StepSecurity attribute the infrastructure overlap to Sapphire Sleet campaigns. Microsoft's report does not itself publish IOCs; network and file indicators below come from the third-party Rust incident analyses and the mapping of that incident to the Microsoft-named ProcmacSteal detections is an analyst inference from the shared August 2026 Rust timeframe.

For operational scaling, actors automate repetitive tasks, summarize victim information and accelerate post-compromise decisions. Microsoft reports Onyx Sleet and Pearl Sleet adopting agentic AI via the open-source ARTEX penetration-testing framework from late October 2026, and Cursor (AI IDE) is named as a development platform.

MITRE ATT&CK techniques used in TL-2026-4755

Defense Evasion

T1027 Obfuscated Files or Information; T1027.004 Obfuscated Files or Information: Compile After Delivery; T1684.001 Impersonation

Execution

T1059.001 Command and Scripting Interpreter: PowerShell

Initial Access

T1195.001 Supply Chain Compromise: Compromise Software Dependencies and Development Tools; T1566 Phishing

Resource Development

T1583.001 Acquire Infrastructure: Domains; T1585.001 Establish Accounts: Social Media Accounts; T1588.007 Obtain Capabilities: Artificial Intelligence

Reconnaissance

T1589 Gather Victim Identity Information; T1591 Gather Victim Org Information; T1592 Gather Victim Host Information; T1593 Search Open Websites/Domains

Affected products and versions in North Korean Threat Actors Operationalize AI Across

  • Rust / crates.io — arrayref
    Vulnerable versions: 0.3.10
    Fixed in: 0.3.9
  • Rust / crates.io — internment
    Vulnerable versions: 0.8.7
    Fixed in: 0.8.6
  • Rust / crates.io — append-only-vec
    Vulnerable versions: 0.1.9
    Fixed in: 0.1.8
  • GitHub — Private and public code repositories (trojanization target)

Remediation for North Korean Threat Actors Operationalize AI Across

Patches

  • Pin arrayref =0.3.9, internment =0.8.6 and append-only-vec =0.1.8

Immediate actions

  • Search Cargo.lock files for proc-macro1, proc-macro-en, aovine, arone, aronenao, tinymember, arrayref 0.3.10, internment 0.8.7 and append-only-vec 0.1.9
  • Remove cached artifacts such as arrayref-0.3.10.crate and proc-macro1-*.crate from ~/.cargo/registry/cache and rebuild affected hosts and CI runners
  • Block 23.254.165.112, 23.254.167.107 and 23.254.167.216 at the perimeter and hunt for outbound connections on ports 443 and 9089
  • Rotate credentials and tokens exposed on any host that built the affected crates

Workarounds

  • Restrict outbound network access from build environments
  • Treat unsolicited recruiter repositories as untrusted and open them only in isolated environments

Longer-term hardening

  • Enable Defender for Office 365 Safe Links, Safe Attachments and zero-hour auto purge
  • Block device code flow in Conditional Access
  • Deploy EDR in block mode and attack surface reduction rules for script execution
  • Vet remote hires: cross-check social media accounts, resumes and references, and verify identity over multiple video sessions
  • Pin and verify dependencies and review build.rs additions to zero-dependency crates

Timeline of North Korean Threat Actors Operationalize AI Across

  • GitHub publishes a security alert on social engineering of technology employees via fake developer and recruiter accounts and malicious repositories, later attributed by Microsoft to Jade Sleet
  • Coral Sleet begins (April 2026 onward) an AI-assisted workflow to trojanize GitHub repositories
  • Jade Sleet starts a job-themed targeting campaign using three front companies with tailored domains, running through 2026-06-12
  • Onyx Sleet generates over 100 target-specific PDF lure files with AI during August-September 2026
  • Attacker-owned crates arone and aronenao receive their final malicious publishes
  • Malicious crates reported to RustSec at 07:54 UTC and deleted from crates.io between 08:03 and 09:25 UTC, an overall exposure window of about 134 minutes
  • Sapphire Sleet-attributed compromise: typosquat proc-macro1 weaponized and arrayref 0.3.10, internment 0.8.7 and append-only-vec 0.1.9 published with the dependency between 07:11 and 07:37 UTC
  • Cloud Security Alliance research note on Jade Sleet macOS backdoors triggered through Cursor AI
  • Microsoft publishes 'Tracking AI adoption in North Korean threat actor tradecraft', noting Onyx Sleet and Pearl Sleet adoption of the ARTEX agentic framework from late October 2026

Sources cited for North Korean Threat Actors Operationalize AI Across

Detection coverage for TL-2026-4755

As of 2026-10-29, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-4755 across Splunk SPL, Microsoft KQL and Sigma, covering 28 indicator(s) of compromise. The whole corpus is readable without an account; a free account unlocks full detection query text in Splunk SPL, Microsoft KQL and Sigma; paid tiers add raw indicator values, correlation and the MCP server. Threadlinqs MCP server · View plans.

9 detection rules (Splunk SPL, Microsoft KQL, Sigma) · Blue and above. Compare plans
28 indicators of compromise · Red and above. Compare plans

Community OSINT corroboration for TL-2026-4755

6 of this threat's indicators have also been reported by the open-source security community. Community sightings are unverified and are kept separate from Threadlinqs' curated indicators. Indicator values, reporters and campaign linkage are available to authenticated Red-tier users.

Threadlinqs Intelligence — Real-Time Threat Detection Platform

[ 0 threats ] [ 0 det ] [ CRIT: 0 ] [ HIGH: 0 ]
// threat_feed
$ sort --newest
Showing all threats

Live intelligence console

Threat level
Fig. 01 · Threat weatherIndexing the archive…
1 square = 1 threat · click to open

Latest Threats