Skip to main content
Mallory
Malware3 associated actors1 associated CVE

Hermes

Tracked since Nov 25, 2025 · Updated today
Trending this weekNot trendingOutside this week’s top 100
Mention activity · 30 days7 mentions · peak 2 in a dayToday
Vulns1Associated
Actors3Associated
ATT&CK46Techniques
IOCs27Tracked
Type
Ransomware
Capabilities
ExtortionPersistenceDefense evasion
Delivery
Exploit kitDrive by
Platforms
Windows

Hermes is a Windows ransomware family first observed in February 2017. It encrypts files on local drives and accessible network resources using AES-256, with RSA protecting the per-file encryption keys. Hermes 2.1 generates a victim-specific RSA key pair and appends an encryption marker and protected key material to encrypted files without changing their names. It leaves ransom instructions and deletes volume shadow copies and backup files to impede recovery.

Hermes 2.1 establishes persistence through the Windows Startup mechanism, relocates and relaunches itself, and deletes its original executable. It repeatedly requests administrative execution of supporting scripts but can continue encrypting files without approval. This version exits on systems configured for Russian, Belarusian, or Ukrainian languages and excludes selected system and application directories from encryption.

Hermes has been delivered by Azorult and through the GreenFlash Sundown exploit kit. A campaign targeting South Korean users exploited Adobe Flash Player vulnerability CVE-2018-4878 through compromised websites to deliver Hermes 2.1. APT38 has used Hermes for file encryption, and Lazarus-linked operations have deployed it after financial intrusions to obstruct investigation, including the 2017 Far Eastern International Bank attack in Taiwan. Hermes was also sold in underground markets, and its codebase was adapted into Ryuk; this code relationship does not establish common operators or developers.

Is this in your environment?See whether this family's indicators match your telemetry and which of your assets are exposed to the CVEs associated with it.
Check your exposure
Associated CVEs

Associated vulnerabilities

1 CVE Mallory has correlated with this family across public research and vendor advisories. Each card links to the full Mallory page for that vulnerability.

CVE-2018-4878via malwarebytes labsUse-after-free remote code execution in Adobe Flash PlayerGreenFlash Sundown had started to use this recent Flash zero-day to distribute the Hermes ransomware. The payload from this attack is Hermes ransomware, version 2.1.View vulnerability →Source ↗
Threat actors

Associated threat actors

3 distinct threat actors Mallory has associated with this family across public reporting. Open in Mallory to see the full evidence chain and overlapping campaigns.

View more details →
APT38

Malware associated with BlueNorOff include: "...Dyepack, Hermes, Twopence, Electricfish..."

via wikipedia enen.wikipedia.org
DPRK

GreenFlash Sundown had started to use this recent Flash zero-day to distribute the Hermes ransomware. The payload from this attack is Hermes ransomware, version 2.1.

via malwarebytes labsmalwarebytes.com
Lazarus

Curiously, our research lead us to connect the nature of Ryuk’s campaign and some of its inner-workings to the HERMES ransomware, a malware commonly attributed to the notorious North Korean APT Lazarus Group.

via checkpoint research blogresearch.checkpoint.com
MITRE ATT&CK

Techniques & procedures

46 distinct techniques documented for this family, organized by ATT&CK tactic.

Reconnaissance

6 techniques
T1589Gather Victim Identity InformationEvidence2

Some endpoints were unauthenticated, allowing the system to retrieve employee information, including names, departments, and SSO account identifiers.

T1590Gather Victim Network InformationEvidence3

This enabled the agents to map 21 connected government systems, including single sign-on infrastructure.

T1593Search Open Websites/DomainsEvidence3

“The attacker used a website traffic-ranking service to find valuable targets in the list Strix produced and prioritized those running custom software.”

T1595Active ScanningEvidence6

The framework used Hermes and OpenClaw, deploying up to 8 sub-agents in parallel to perform reconnaissance, credential attacks, API testing, data collection, and lateral movement. | Researchers found evidence that the agents discovered hidden API endpoints on a government web application that returned valid authenticated sessions without requiring credentials.

T1595.001Scanning IP BlocksEvidence1

These agents mapped 21 government systems, researched vulnerabilities, adapted tactics whenever blocked, and moved through the network with minimal human steering.

T1595.002Vulnerability ScanningEvidence2

These are autonomous sessions where the models search vulnerability databases, GitHub repositories, and other security research for specific techniques, CVEs, and common weaknesses to exploit in the targeted government's infrastructure.

Initial Access

3 techniques
T1078Valid AccountsEvidence12

A human operator pasted 301 results into the console, but also handpicked at least two targets for which they already had an administrator password.

T1133External Remote ServicesEvidence1

AI agents can rapidly chain multiple attack methods and use secondary systems — such as backup or testing environments — as springboards

T1190Exploit Public-Facing ApplicationEvidence10

The generated reports were then handed to the autonomous penetration testing engine Cairn, which was used to launch 105 attack projects... Each attack path was chosen by the harness in real time through extensive probing and exploitation attempts.

Execution

3 techniques
T1053.003CronEvidence3

А иногда атакующий менял конфигурацию развертываний Kubernetes и создавал cron-задачи, которые восстанавливали веб-скиммер после удаления.

T1059Command and Scripting InterpreterEvidence4

The attackers enabled the software's so-called "YOLO mode," allowing it to execute commands without waiting for human approval.

T1203Exploitation for Client ExecutionEvidence2

Hermes ransomware distributed to South Koreans via recent Flash zero-day

Persistence

4 techniques
T1053.003CronEvidence3

А иногда атакующий менял конфигурацию развертываний Kubernetes и создавал cron-задачи, которые восстанавливали веб-скиммер после удаления.

T1078Valid AccountsEvidence12

A human operator pasted 301 results into the console, but also handpicked at least two targets for which they already had an administrator password.

T1133External Remote ServicesEvidence1

AI agents can rapidly chain multiple attack methods and use secondary systems — such as backup or testing environments — as springboards

T1505.003Web ShellEvidence4

The agent used an SQL injection to obtain a plaintext one-time password, access a web panel, upload a web shell...

Privilege Escalation

4 techniques
T1053.003CronEvidence3

А иногда атакующий менял конфигурацию развертываний Kubernetes и создавал cron-задачи, которые восстанавливали веб-скиммер после удаления.

T1068Exploitation for Privilege EscalationEvidence6

What once required teams of skilled operators working in shifts can now be orchestrated by software that maps networks, steals credentials, discovers flaws, and pivots in real time.

T1078Valid AccountsEvidence12

A human operator pasted 301 results into the console, but also handpicked at least two targets for which they already had an administrator password.

T1548.003Sudo and Sudo CachingEvidence1

The agent ... upload[ed] a web shell, escalate[d] privileges through a misconfigured sudo rule, and access[ed] AWS credentials.

Stealth

4 techniques
T1070Indicator RemovalEvidence1

Once the theft is complete, they will try to destroy all evidence by deploying crimeware ransomware or wipers.

T1070.004File DeletionEvidence2

The cybersecurity firm also identified a Hermes agent skill designed to delete the stolen card data from the victim’s Magento database. Additionally, the agent deleted a bicycle retailer’s backup tables after being instructed to erase staging tables created within the database.

T1078Valid AccountsEvidence12

A human operator pasted 301 results into the console, but also handpicked at least two targets for which they already had an administrator password.

T1684.001ImpersonationEvidence1

attackers used standard social engineering tactics to bypass the guardrails of unidentified frontier models to create a penetration testing framework to attack a government agency

Defense Impairment

1 technique
T1578Modify Cloud Compute InfrastructureEvidence1

В разных случаях злоумышленник... модифицировал содержимое S3 и CDN, серверный кеш и поля БД.

Credential Access

5 techniques
T1056.003Web Portal CaptureEvidence2

ИИ-агенты внедряли на сайты магазинов веб-скиммеры... добавлял вредоносный код в легитимные JavaScript-файлы и на страницы оплаты, встраивал его в блоки Google Tag.

T1110.003Password SprayingEvidence3

The framework also conducted automated password spraying against an office automation portal. It used employee usernames collected from exposed APIs and solved CAPTCHA images with OCR. By testing predictable password patterns, the agents cracked 85 accounts across several rounds.

T1539Steal Web Session CookieEvidence1

Session cookies : PHPSESSID, XSRF-TOKEN, session Laravel du panneau admin ICT du MOF

T1621Multi-Factor Authentication Request GenerationEvidence1

Using employee usernames harvested from an unauthenticated API, the agents broke into a government department’s office automation portal, solving its CAPTCHAs with 100 percent accuracy.

T1649Steal or Forge Authentication CertificatesEvidence3

A government API reportedly accepted tokens with the none algorithm, allowing attackers to forge authentication tokens without access to a signing key.

Discovery

8 techniques
T1007System Service DiscoveryEvidence2

Five recovered Hermes call logs show the agent was used to find a way to elevate privileges, scan for kernel vulnerabilities, enumerate services...

T1046Network Service DiscoveryEvidence4

The researchers say the attackers assembled an autonomous hacking platform using open-source AI-agent frameworks, enabling multiple agents to simultaneously map networks, research vulnerabilities, attempt intrusions, and adapt tactics when an attack path failed.

T1069Permission Groups DiscoveryEvidence1

Five recovered Hermes call logs show the agent was used to... search for SUID and SGID binaries...

T1082System Information DiscoveryEvidence5

Threat actors used it to perform system enumeration, escalate privileges, discover files and services, and conduct network reconnaissance.

T1083File and Directory DiscoveryEvidence9

Threat actors used it to perform system enumeration, escalate privileges, discover files and services, and conduct network reconnaissance.

T1135Network Share DiscoveryEvidence1

Additional logs indicate the operator instructed the agent to enumerate a content directory containing PDF, DOC, XLS files, and personnel records associated with the Office of Permanent Secretary for Finance.

T1518Software DiscoveryEvidence2

It extracted API endpoints, OAuth client IDs, Keycloak configuration data, and authentication details.

T1613Container and Resource DiscoveryEvidence1

Five recovered Hermes call logs show the agent was used to... inspect containers...

Lateral Movement

1 technique
T1021Remote ServicesEvidence1

And then, the agents pivoted to the Taiwanese government’s supply chain.

Collection

4 techniques
T1056.003Web Portal CaptureEvidence2

ИИ-агенты внедряли на сайты магазинов веб-скиммеры... добавлял вредоносный код в легитимные JavaScript-файлы и на страницы оплаты, встраивал его в блоки Google Tag.

T1119Automated CollectionEvidence1

The platform continuously assessed available evidence, ranked possible attack paths, and reprioritized them as circumstances changed.

T1213Data from Information RepositoriesEvidence12

После кражи и скачивания данных банковских карт агент должен был удалить их из базы Magento.

T1560Archive Collected DataEvidence1

Hermes изучил веб-каталог центрального аппарата Минфина и составил список PDF-, DOC- и XLS-файлов

Command and Control

5 techniques
T1071Application Layer ProtocolEvidence2

The agents are bridged to Telegram so the operator can task them from a phone and collect results in a group chat.

T1071.001Web ProtocolsEvidence1

155.254.22.215 Staging and command server, AI console... 213.21.239.62 C2.

T1090ProxyEvidence1

They then pasted 301 results into the console with a message that ended with 跑这些 用代理 只扫高危 (“run these, use the proxy, high severity only”).

T1102Web ServiceEvidence1

La liste des TTPs identifie T1102 ; les opérateurs utilisent OpenRouter pour accéder aux modèles IA utilisés dans la campagne.

T1105Ingress Tool TransferEvidence2

Azorult can download and execute additional files. Azorult has also downloaded a ransomware payload called Hermes.

Exfiltration

1 technique
T1041Exfiltration Over C2 ChannelEvidence3

The campaign stole “over 600,000 credit card records” and a Hermes skill was titled “Database Wipe After Extraction.”

Impact

4 techniques
T1485Data DestructionEvidence6

После кражи и скачивания данных банковских карт агент должен был удалить их из базы Magento. В результате несколько пострадавших магазинов столкнулись с потерей данных и сбоями в работе.

T1486Data Encrypted for ImpactEvidence7

Tracking Ransomware End-to-end Cerber Locky WannaCryptor

T1490Inhibit System RecoveryEvidence1

Finally, Ryuk will destroy its encryption key and execute a .BAT file that will delete shadow copies and various backup files from the disk.

T1565.001Stored Data ManipulationEvidence1

“After extracting and downloading all card data, wipe the source fields in batches.”

Indicators of compromise

Observables

27 indicators attributed across vendor reports, sandbox runs, and researcher write-ups. Values are masked here and available in full in Mallory.

View all in app →
Indicator values are available in Mallory. Pivot on every hash, domain, URL, and IP, or pipe them straight into your SIEM.View all in Mallory
Activity feed

Recent activity

34 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.

View more in app →
+24 more in the timeline. Sign up to see them
Is this in your environment?See whether this family's indicators match your telemetry and which of your assets are exposed to the CVEs associated with it.
Check your exposure
Mallory platform

The version that knows your environment.

This page is what’s public. Mallory adds the parts that aren’t: which of your assets match these IOCs, which detections are missing, which campaigns to expect next, and what to do in the next 30 minutes.
IOC matching27Match every observed IP, domain, and hash against your live telemetry.
Associated threat actors3Groups linked to this family, with evidence pinned to each claim.
Associated vulnerabilities1CVEs linked to this family, with the reporting behind each link.
Detection signaturesYARA, Sigma, Snort, and vendor rules, auto-deployed to your SIEM.
MITRE ATT&CK mapping46Every documented technique, ranked by evidence weight.
Researcher chatterReddit, Mastodon, and CTI community discussion around this family.