KawaiiGPT
- Capabilities
- Credential theftLateral movementReconnaissanceExfiltration
- Platforms
- LinuxAndroidWindows
KawaiiGPT is a free, open-source offensive AI assistant first identified in July 2025 and distributed through GitHub. Marketed as an unrestricted, anime-themed alternative to safeguarded AI assistants, it provides a conversational interface for generating malicious code and social-engineering material. It is an attack-enablement tool rather than a conventional malware payload. The tool runs on Linux and supports Android through Termux.
Demonstrated functionality includes generating bank-impersonation spearphishing emails that direct recipients to credential- and financial-information-harvesting websites. It can also generate Python scripts for lateral movement on Linux using Paramiko to authenticate over SSH and establish remote shells. Additional demonstrated output includes Windows-targeted scripts that locate email files and exfiltrate them as email attachments. These capabilities assist with credential theft, lateral movement, reconnaissance, and data exfiltration; they represent generated attack tooling rather than evidence that the assistant independently performs those actions. Its free availability and conversational workflow reduce the programming expertise required to develop offensive scripts.
Techniques & procedures
20 distinct techniques documented for this family, organized by ATT&CK tactic.
Reconnaissance
1 technique
Reconnaissance
key capabilities have been segmented into phishing automation, malware development, reconnaissance, brute force, vulnerability exploitation, and social engineering.
Resource Development
4 techniques
Resource Development
Malware (T1587.001) — разработка AI-малвари и вредоносных скриптов с помощью LLM
Artificial Intelligence (T1588.007) — приобретение доступа к AI-capabilities для подготовки атаки
Generate Content (T1683) — использование AI для генерации контента атаки
Written Content (T1683.001) — генерация фишинговых писем, BEC-сообщений, scam-скриптов
Initial Access
4 techniques
Initial Access
The resulting script does not introduce hugely novel capabilities, but it automates a standard, critical step in nearly every successful breach... as the generated code “authenticates as a legitimate user and grants the attacker a remote shell onto the new target machine.”
AI has made significantly easier... Drafting phishing lures... Threat actors use it to systematically design lookalike phishing pages... draft convincing spear-phishing lures.
Threat actors are adopting artificial intelligence to reduce time spent on attack vectors and improve the quality of attacks... Its application ranges from generating spear-phishing emails to lure images and dynamic payload generation executed in real time.
The researchers tested its capabilities using prompts instructing it to create: a spear-phishing message generation with realistic domain spoofing and credential-harvesting links.
Execution
1 technique
Execution
a Python script for lateral movement that used the paramiko SSH library ... a Python script that recursively looked on a Windows filesystem for target files using os.walk, and then used Python’s smtplib library to pack and exfiltrate the data
Persistence
1 technique
Persistence
The resulting script does not introduce hugely novel capabilities, but it automates a standard, critical step in nearly every successful breach... as the generated code “authenticates as a legitimate user and grants the attacker a remote shell onto the new target machine.”
Privilege Escalation
1 technique
Privilege Escalation
The resulting script does not introduce hugely novel capabilities, but it automates a standard, critical step in nearly every successful breach... as the generated code “authenticates as a legitimate user and grants the attacker a remote shell onto the new target machine.”
Stealth
2 techniques
Stealth
The resulting script does not introduce hugely novel capabilities, but it automates a standard, critical step in nearly every successful breach... as the generated code “authenticates as a legitimate user and grants the attacker a remote shell onto the new target machine.”
key capabilities have been segmented into phishing automation, malware development, reconnaissance, brute force, vulnerability exploitation, and social engineering.
Credential Access
1 technique
Credential Access
Prompts produce credential-harvesting emails with professional formatting and subject lines such as "Urgent: Verify Your Account Information."
Discovery
1 technique
Discovery
a Python script that recursively looked on a Windows filesystem for target files using os.walk
Lateral Movement
2 techniques
Lateral Movement
a Python script for lateral movement that used the paramiko SSH library to connect to a host and execute commands remotely via exec_command()
Sellers tout benefits such as... Python script for SSH lateral movement with remote shell access in under a minute.
Collection
2 techniques
Collection
Prompts produce credential-harvesting emails with professional formatting and subject lines such as "Urgent: Verify Your Account Information."
Data exfiltration prompts generate code that goes through every folder and subfolder to find EML files - the standard format for saved emails - and then uses Python's built-in email-sending tool smtplib to forward those files to an attacker.
Command and Control
2 techniques
Command and Control
the researchers warn that its command execution capability could allow attackers to escalate privileges, steal data, and drop and execute additional payloads.
Lateral movement requests generate Python scripts using paramiko for SSH authentication and remote shell access.
Exfiltration
2 techniques
Exfiltration
The script then sent the stolen files as email attachments to an attacker-controlled address.
used Python’s smtplib library to pack and exfiltrate the data to an attacker-controlled address
Observables
2 indicators attributed across vendor reports, sandbox runs, and researcher write-ups. Values are masked here and available in full in Mallory.
Recent activity
11 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.