ShadowRay 2.0
- Type
- Botnet
- Capabilities
- Credential theftExfiltrationPersistenceInitial accessLateral movementDefense evasionReconnaissanceScanningDDOS
- Platforms
- LinuxCloud
ShadowRay 2.0 is a self-propagating, multipurpose botnet operation targeting exposed Ray distributed-computing clusters, particularly Linux-based AI infrastructure with valuable CPU and NVIDIA GPU resources. Documented in November 2025, the operation has been active since at least September 2024. It was initially attributed to an operator tracked as IronErn440 and subsequently linked to TeamPCP.
The infection chain exploits unauthenticated job submission associated with CVE-2023-48022 to execute multistage Bash and Python payloads through exposed Ray dashboards. Ray's maintainers dispute the vulnerability classification because the framework assumes deployment in a trusted network. Malicious jobs enumerate available resources and abuse Ray's orchestration and node-affinity scheduling to execute across cluster nodes, including nodes not directly exposed to the internet. Compromised systems also scan for additional exposed Ray clusters and propagate autonomously.
The operation deploys cryptocurrency miners, including XMRig, and supports cloud and database credential theft, theft of AI models, datasets and source code, interactive control through Python reverse shells, and DDoS attacks using Sockstress. It targets organizations globally, emphasizing large clusters and GPU-rich environments rather than a single industry.
Payloads are distributed and updated through disposable GitLab and GitHub repositories. Persistence uses scheduled tasks and systemd modifications, including periodic retrieval of updated payloads. Defense-evasion measures include masquerading as legitimate system processes, concealing payloads, and limiting CPU consumption to approximately 60 percent. The malware also terminates competing miners and blocks access to competing mining pools to monopolize compromised computing resources.
Associated vulnerabilities
2 CVEs Mallory has correlated with this family across public research and vendor advisories. Each card links to the full Mallory page for that vulnerability.
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.
«Принципиальное отличие: ShadowRay 2.0 - самовоспроизводящийся ботнет. Скомпрометированные Ray-ноды автоматически ищут другие доступные кластеры и распространяют инфекцию.»
The firm also assessed that TeamPCP ran ShadowRay 2.0, the campaign it documented in November 2025 against exposed Ray clusters, which it had attributed at the time to an actor called IronErn440.
The firm also assessed that TeamPCP ran ShadowRay 2.0, the campaign it documented in November 2025 against exposed Ray clusters, which it had attributed at the time to an actor called IronErn440.
Techniques & procedures
3 distinct techniques documented for this family, organized by ATT&CK tactic.
Initial Access
1 technique
Initial Access
«Initial Access T1190 (Exploit Public-Facing Application) ... POST к /api/jobs/ на Ray Dashboard»
Impact
2 techniques
Impact
«Compute Hijacking (T1496.001, Impact) ... запуск XMRig и других майнеров на GPU»
AI-Generated Malware Attacks 230,000 Exposed Ray AI Clusters in Massive ShadowRay 2.0 Botnet Campaign ... Ddos
Observables
5 indicators attributed across vendor reports, sandbox runs, and researcher write-ups. Values are masked here and available in full in Mallory.
Recent activity
12 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.