Bizarre Bazaar Campaign Hijacks Exposed LLM Endpoints for Cybercrime
- Securityish
- Threats & Incidents
Quick Summary
A malicious campaign named ‘Bizarre Bazaar’ is targeting exposed Large Language Model (LLM) service endpoints, resulting in over 35,000 recorded attack sessions. Researchers from Pillar Security discovered that attackers exploit weakly protected AI infrastructure to steal computing resources, resell API access, and exfiltrate sensitive data. This ongoing operation highlights the risks associated with misconfigured AI endpoints, which can lead to significant costs and data exposure for organizations.
The Securityish Brief
Pillar Security has identified a large-scale cybercrime operation called ‘Bizarre Bazaar’ that targets exposed LLM service endpoints. Over 40 days, they recorded more than 35,000 attack sessions on their honeypots, revealing a systematic approach to exploiting poorly authenticated AI infrastructure. The attackers aim to monetize unauthorized access by stealing computing resources for cryptocurrency mining, reselling API access on darknet markets, and exfiltrating sensitive data from prompts and conversation history.
The campaign is notable for being one of the first attributed examples of ‘LLMjacking’ attacks, which differ from traditional API abuse due to the significant costs associated with compromised LLM endpoints. Attackers typically exploit misconfigurations, such as unauthenticated Ollama endpoints on port 11434 and OpenAI-compatible APIs on port 8000. These attacks often commence within hours of a misconfigured endpoint being detected in internet scans.
Understanding the Threat Landscape
The Bizarre Bazaar operation involves a criminal supply chain with three distinct threat actors. The first actor uses bots to scan the internet for LLM and Model Context Protocol (MCP) endpoints, while the second validates these findings. The third actor operates a commercial service called ‘silver[.]inc,’ which resells access to these endpoints in exchange for cryptocurrency or PayPal payments.
SilverInc promotes a project named NeXeonAI, marketed as a unified AI infrastructure that provides access to over 50 AI models from leading providers. This operation continues to pose risks as it allows attackers to pivot into internal systems via MCP servers, which can lead to further exploitation.
As of now, the Bizarre Bazaar campaign remains active, and the SilverInc service continues its operations. The ongoing nature of this threat underscores the importance of securing AI infrastructure against unauthorized access and exploitation.
Key Takeaways
- Regularly audit your AI infrastructure for misconfigurations and ensure all endpoints are properly secured.
- Implement strong authentication mechanisms for all LLM and MCP endpoints to prevent unauthorized access.
- Monitor for unusual activity or access patterns that may indicate exploitation of your AI services.
- Stay informed about emerging threats and tactics used by cybercriminals targeting AI technologies.
- Consider using security tools that can detect and alert on exposed endpoints in real-time.
Key Terms & Concepts
- LLM: In this article, LLM refers to Large Language Models, which are AI systems designed to understand and generate human-like text.
- LLMjacking: LLMjacking describes a type of cyber attack that exploits vulnerabilities in LLM endpoints for unauthorized access and monetization.
- MCP: Model Context Protocol (MCP) is a communication protocol used by AI models to interact with other systems and services.
- SilverInc: SilverInc is a commercial service involved in reselling access to compromised AI infrastructure in exchange for cryptocurrency or PayPal payments.
Read the original article here: https://www.bleepingcomputer.com/news/security/hackers-hijack-exposed-llm-endpoints-in-bizarre-bazaar-operation/
Your 5-Minute Securityish Brief
A weekly digest of cybersecurity news, phishing alerts, privacy tips, and emerging threats, simplified so anyone can understand what matters and why.
Securityish
Securityish explains cybersecurity, scams, data breaches, and privacy risks in simple language so you know what’s happening and how to protect yourself.
© 2026 Securityish
