Cybercriminals Exploit AI Developer Tools in Sophisticated Malware Campaign
Cybercriminals are targeting AI developers by cloning trusted GitHub repositories and embedding malware in fake developer tools, according to Netskope Threat Labs. The campaign, linked to a Malware-as-a-Service (MaaS) operation, previously used ClickFix social-engineering tactics to distribute a Node.js-based infostealer.
Attackers lure victims searching for AI resources, coding assistants, Python security guides, Rust frameworks, Claude-related tools, and ComfyUI projects. They replicate legitimate repositories, maintaining authentic appearances while altering installation links or adding malicious files in seemingly harmless folders. This tactic is particularly dangerous as developers often have access to sensitive assets like source code, cloud environments, API keys, and CI/CD systems. A compromised credential could enable attackers to infiltrate an organization’s software supply chain.
Victims have been identified across North America, Asia, and Southern Europe, with financial services, banking, and technology sectors among the hardest hit. The malware is typically delivered as a ZIP archive containing four files: lua51.dll (a legitimate LuaJIT runtime), compiler.exe (a renamed LuaJIT interpreter), Application.bat, and gc.txt. The batch file executes the interpreter, which runs gc.txt an obfuscated Lua script that initiates the infection.
A key feature of this campaign is EtherHiding, a technique that uses blockchain smart contracts to dynamically retrieve command-and-control (C2) server addresses. Instead of hardcoding C2 domains, the malware queries a Polygon smart contract via public RPC providers, allowing attackers to redirect infected systems by updating the contract. This evasion tactic enhances the malware’s persistence.
The first-stage loader downloads a second Lua payload (dist.lua) alongside additional runtime files. Researchers identified GitHub accounts, including yawalinte and JuliusMAAR, hosting identical second-stage payloads before takedowns. The second stage, obfuscated with MoonSec V3, repeats geolocation checks, encrypts communications with an XOR key, and employs EtherHiding with fallback RPC providers for resilience.
The final payload includes multiple infostealer variants, including the previously analyzed NodeJS-based MaaS malware. The campaign underscores the risks of unverified GitHub repositories and the growing sophistication of supply chain attacks targeting developers.
Source: https://cyberpress.org/malicious-ai-tools-exposed/
GitHub TPRM report: https://www.rankiteo.com/company/github
"id": "git1785839378",
"linkid": "github",
"type": "Cyber Attack",
"date": "8/2026",
"severity": "85",
"impact": "4",
"explanation": "Attack with significant impact with customers data leaks"{'affected_entities': [{'industry': ['Financial services',
'Banking',
'Technology'],
'location': ['North America',
'Asia',
'Southern Europe'],
'type': 'Organization'}],
'attack_vector': ['Cloned GitHub repositories',
'Malicious developer tools',
'Social engineering'],
'data_breach': {'data_exfiltration': True,
'sensitivity_of_data': 'High',
'type_of_data_compromised': ['Source code',
'API keys',
'Cloud environment credentials']},
'description': 'Cybercriminals are targeting AI developers by cloning trusted '
'GitHub repositories and embedding malware in fake developer '
'tools, according to Netskope Threat Labs. The campaign, '
'linked to a Malware-as-a-Service (MaaS) operation, previously '
'used ClickFix social-engineering tactics to distribute a '
'Node.js-based infostealer. Attackers lure victims searching '
'for AI resources, coding assistants, Python security guides, '
'Rust frameworks, Claude-related tools, and ComfyUI projects. '
'They replicate legitimate repositories, maintaining authentic '
'appearances while altering installation links or adding '
'malicious files in seemingly harmless folders. This tactic is '
'particularly dangerous as developers often have access to '
'sensitive assets like source code, cloud environments, API '
'keys, and CI/CD systems. A compromised credential could '
'enable attackers to infiltrate an organization’s software '
'supply chain.',
'impact': {'data_compromised': ['Source code',
'Cloud environments',
'API keys',
'CI/CD systems'],
'operational_impact': 'Potential infiltration of organizational '
'software supply chain',
'systems_affected': ['Developer workstations',
'Software supply chain']},
'initial_access_broker': {'entry_point': 'Cloned GitHub repositories',
'high_value_targets': ['Developers with access to '
'sensitive assets']},
'lessons_learned': 'The campaign underscores the risks of unverified GitHub '
'repositories and the growing sophistication of supply '
'chain attacks targeting developers.',
'motivation': ['Data exfiltration', 'Supply chain infiltration'],
'post_incident_analysis': {'root_causes': ['Use of unverified GitHub '
'repositories',
'Malicious files in cloned '
'repositories']},
'references': [{'source': 'Netskope Threat Labs'}],
'response': {'third_party_assistance': 'Netskope Threat Labs'},
'threat_actor': 'Malware-as-a-Service (MaaS) operation',
'title': 'Cybercriminals Exploit AI Developer Tools in Sophisticated Malware '
'Campaign',
'type': 'Malware Campaign',
'vulnerability_exploited': 'Unverified GitHub repositories'}