Threat reportMalwareTL-2026-4849

APT10 LODEINFO Malware Evolution: v0.6.6 - v0.7.3 Analysis

highACTIVE

APT10 LODEINFO Malware Evolution (TL-2026-4849), also tracked as LODEINFO, is a high-severity malware campaign, first published 2026-10-30. It is attributed to APT10 (China) with medium confidence, affects Microsoft Microsoft Word / Office (macro and remote template abuse), maps to 15 MITRE ATT&CK techniques (T1027, T1055, T1056.001), and is covered by 9 detection rules and 25 indicators of compromise.

Severity
HIGHAssessed severity
CVEs
0None referenced
Techniques
15MITRE ATT&CK
Actors
1APT10
Detection rules
9SPL · KQL · Sigma
IOCs
25Indicators of compromise

Key facts for TL-2026-4849

Threat ID
TL-2026-4849
Also known as
LODEINFO, DOWNIISSA
Severity
HIGH
Status
ACTIVE
Category
MALWARE
First published
Last reviewed
Attribution
APT10
Attribution confidence
MEDIUM
Nation-state nexus
China
Motivation
ESPIONAGE
Target sectors
news - media, government administration, diplomacy, defense, think-tank
Target regions
japan
Detection rules
9
Indicators of compromise
25

Malware and tooling in APT10 LODEINFO Malware Evolution

Malware and tooling: DOWNIISSA - S9021, LODEINFO - S9020

How APT10 LODEINFO Malware Evolution works

ITOCHU Cyber & Intelligence analyzes LODEINFO fileless backdoor versions v0.6.6 through v0.7.3, attributed to APT10 and used against Japanese media, diplomatic, public-sector, defense-industry and think-tank targets. Techniques include remote template injection, DLL side-loading (Elze.exe / Frau.dll), control-flow flattening and in-memory deployment.

LODEINFO is a fileless backdoor delivered by spear-phishing since December 2019 and first named by JPCERT/CC in February 2020. ITOCHU Cyber & Intelligence (published 2024-01-24) documents versions v0.6.6 through v0.7.3. The infection chain starts with a malicious Word attachment whose macro selects 32-bit or 64-bit downloader shellcode according to OS architecture. From v0.6.9 the document uses remote template injection: Word fetches a template (for example https://45.76.222.130/template.dotm) from attacker infrastructure, and the template's VBA carries the shellcode, which keeps the initial document cleaner and harder to detect.

The decrypted payload contains three objects: a legitimate host executable (Elze.exe), a malicious side-loaded library (Frau.dll) and an XOR-encoded LODEINFO shellcode blob (Elze.exe_bak). Frau.dll decodes the blob and injects the backdoor into memory, so LODEINFO runs without a file artifact. Obfuscation includes control-flow flattening, junk code, Base64-fragmented VBA shellcode reassembled at run time, and self-patching shellcode that decodes itself with an incrementing XOR key (0x00-0xFF). Version 0.7.0 changed the API-name hashing logic and uses a per-sample hardcoded XOR key. Version 0.7.0 also added a check for Japanese Microsoft Office language settings, which v0.7.1 removed, indicating possible expansion beyond Japanese-language environments. Filenames shifted from Japanese to English.

Version 0.7.1 introduced an intermediate stage disguised as a PEM file. It is decoded by stripping the header and footer, Base64-decoding, removing a 3-byte header, deriving a key with HMAC-SHA1 from a hardcoded password, decrypting with AES and applying a single-byte XOR. The backdoor then loads directly in memory. C2 traffic uses multiple Vigenere cipher layers and command IDs hidden with a 2-byte XOR. The backdoor exposes about 18-22 commands (command, ls, rm, mv, cp, cat, mkdir, send, recv, memory, kill, cd, ver, print, ransom, comc, config, runas, keylog, ps, pkill, autorun). keylog, ps, pkill and autorun were restored in v0.7.2/v0.7.3, runas was added in v0.7.1 and config was implemented in v0.7.2. keylog writes keystrokes and timestamps to %temp%%hostname%.tmp.

The downloader shares traits with DOWNIISSA (self-patching, C2 data encoding, decrypted data structure) previously documented by Kaspersky in 2022; earlier Kaspersky analysis also described K7Security-binary DLL side-loading and process injection into msiexec.exe. ITOCHU attributes the activity to APT10 on the basis of TTP similarity and malware relationships; infrastructure favors AS-CHOOPA (Vultr) hosting and was consistent across 2023 versions. Trend Micro tracks overlapping activity as Earth Kasha and cautions that evidence for definitive APT10 attribution is limited, using an 'APT10 Umbrella' term for related intrusion sets; it additionally reports a 2023-2024 campaign using public-facing application exploits to deploy LODEINFO, NOOPDOOR and Cobalt Strike (not part of the ITOCHU analysis). ITOCHU recommends endpoint products capable of in-memory scanning, since both downloader and backdoor are fileless.

MITRE ATT&CK techniques used in TL-2026-4849

Defense Evasion

T1027 Obfuscated Files or Information; T1055 Process Injection; T1140 Deobfuscate/Decode Files or Information; T1221 Template Injection; T1574.001 DLL; T1620 Reflective Code Loading

Collection

T1056.001 Keylogging; T1113 Screen Capture

Discovery

T1057 Process Discovery; T1614.001 System Language Discovery

Execution

T1059.005 Visual Basic; T1204.002 Malicious File

Command and Control

T1071.001 Web Protocols; T1573.001 Symmetric Cryptography

Initial Access

T1566.001 Spearphishing Attachment

Affected products and versions in APT10 LODEINFO Malware Evolution

  • Microsoft — Microsoft Word / Office (macro and remote template abuse)
    Vulnerable versions: Configurations permitting macros and remote template loading
  • Microsoft — Windows (32-bit and 64-bit)
    Vulnerable versions: Hosts running LODEINFO-targeted Office installations

Remediation for APT10 LODEINFO Malware Evolution

Immediate actions

  • Block the listed C2 IPs (167.179.106.224, 167.179.77.72, 172.104.112.218, 202.182.116.25, 45.76.197.236, 45.76.222.130, 45.77.183.161) at the perimeter and hunt historical connections
  • Search endpoints for Elze.exe paired with Frau.dll and for the listed MD5 hashes
  • Quarantine inbound Word documents with macros or remote template references (.dotm) pending review

Workarounds

  • Disable VBA macros by Group Policy for users who do not require them
  • Restrict Office outbound network access to trusted destinations

Longer-term hardening

  • Deploy EDR/endpoint tooling with in-memory scanning to detect fileless shellcode
  • Enforce Microsoft Office policy blocking macros from the internet and external template loading
  • Monitor for DLL side-loading of legitimate signed executables from user-writable paths

Timeline of APT10 LODEINFO Malware Evolution

  • Spear-phishing campaigns with malicious Word attachments delivering LODEINFO begin (December 2019, per ITOCHU).
  • JPCERT/CC first names LODEINFO in a blog post and attributes it to APT10 (February 2020).
  • Politician-themed self-extracting archive deploying LODEINFO observed (per Kaspersky); DOWNIISSA downloader also discovered around June 2022.
  • Kaspersky publishes initial documentation of LODEINFO 2022 activity and DOWNIISSA (November 2022).
  • LODEINFO v0.7.0 adds a Japanese Office language check and new API-name hashing with per-sample XOR keys (around June 2023).
  • Language check removed in v0.7.1 and fake PEM intermediate stage introduced, indicating targeting beyond Japanese-language environments (around July 2023).
  • Versions v0.7.2 and v0.7.3 observed restoring keylog, ps, pkill and autorun commands and implementing config (late 2023 per ITOCHU).
  • ITOCHU Cyber & Intelligence publishes analysis of LODEINFO v0.6.6 through v0.7.3 with IOCs.

Sources cited for APT10 LODEINFO Malware Evolution

Detection coverage for TL-2026-4849

As of 2026-10-30, Threadlinqs Intelligence publishes 9 detection rule(s) for TL-2026-4849 across Splunk SPL, Microsoft KQL and Sigma, covering 25 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
25 indicators of compromise · Red and above. Compare plans

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