Security

AI Infrastructure Security: Securing Gateways and Control Points

2 min read

Summary

Microsoft Security Research warns that AI gateways, retrieval platforms, and orchestration services are emerging as high-value attack targets. Based on observed compromises involving LiteLLM, RAGFlow, and Kestra, the report highlights how attackers are using these systems to steal secrets, gain persistence, and monetize compute—making stronger controls and monitoring essential for defenders.

Need help with Security?Talk to an Expert

Introduction

AI infrastructure is quickly becoming part of the enterprise control plane. Services such as model gateways, retrieval platforms, and workflow orchestrators often sit between users, apps, data, and AI models—meaning they can hold sensitive credentials, configuration data, and execution privileges. Microsoft’s latest research shows attackers are increasingly targeting these layers directly.

What Microsoft observed

Microsoft analyzed three separate compromises affecting AI-related workloads:

  • LiteLLM gateway: Attackers likely exploited exposed gateway surfaces tied to public vulnerability chains, then harvested environment secrets, accessed database-related data, and deployed miner payloads.
  • RAGFlow deployment: Activity included SSRF-style reconnaissance, code execution, and modifications designed to intercept newly configured LLM provider credentials.
  • Kestra workflow environment: Attackers used workflow-origin shell execution, explored Docker and container environments, collected secrets, and monetized compute resources with XMRig.

Across all three cases, the attacker goals were similar:

  • Steal credentials and API keys
  • Establish persistence
  • Access downstream systems and data
  • Abuse compute resources for cryptomining

Why this matters for defenders

The key takeaway is that AI infrastructure now functions as a high-trust control point. A compromise in one of these services can expose:

  • Model-provider API keys
  • Database connection strings
  • Tenant and routing configuration
  • Workflow execution paths
  • Container or host-level access

For IT and security teams, that means AI components should be treated like other critical enterprise infrastructure—not experimental side systems.

Microsoft’s guidance points to several immediate steps:

  • Inventory exposed AI management surfaces across gateways, orchestration tools, and retrieval services.
  • Restrict administrative access using least privilege and network segmentation.
  • Monitor for gateway-originated execution and unusual shell, Python, or container activity.
  • Protect secrets by reducing reliance on environment variables where possible and rotating exposed credentials quickly.
  • Patch vulnerable platforms and review public CVEs affecting exposed AI services.
  • Watch for secret access and outbound exfiltration from AI runtime contexts.

Next steps

Organizations deploying AI services should review whether gateways and orchestration layers are internet-exposed, overprivileged, or lightly monitored. If your team uses LiteLLM, RAGFlow, Kestra, or similar tooling, now is a good time to validate patch levels, audit credential storage, and strengthen detection for suspicious activity originating from AI workloads.

Need help with Security?

Our experts can help you implement and optimize your Microsoft solutions.

Talk to an Expert

Stay updated on Microsoft technologies

AI securityLiteLLMRAGFlowKestraMicrosoft Security

Related Posts

Security

Microsoft Sentinel Named a 2026 SIEM Leader

Microsoft has been named a Leader in the 2026 Gartner Magic Quadrant for SIEM, highlighting the continued momentum behind Microsoft Sentinel. The announcement also underscores Microsoft's push toward integrated, AI-powered security operations through deep Sentinel integration with the new Integrated Security Operations Center in Microsoft Defender.

Security

Microsoft Perception GA: Agentic Security at Ignite

Microsoft has made Microsoft Perception generally available, introducing an agentic security platform that coordinates red, blue, and green team agents through automated playbooks. Announced at Ignite 2026, the update also expands Defender and Sentinel capabilities to help security teams detect, investigate, and remediate AI-speed threats while keeping administrators in control.

Security

Device Code Phishing Risks for AI Coding Agents

Microsoft warns that device code phishing can be used against AI coding agents such as GitHub Copilot CLI, not just user mailboxes or cloud apps. In a controlled lab, a single approved device code sign-in could enable remote session steering, memory poisoning, synced session exposure, and private repository access depending on account permissions and enabled features.

Security

Microsoft 365 Outage Phishing Campaign Targets Users

Microsoft Threat Intelligence reported a high-volume phishing campaign that impersonates Microsoft 365 service-health notifications to trick users into clicking malicious links. The campaign matters because it uses outage-themed urgency and recipient-specific tracking to target multiple industries, increasing the risk of credential theft and follow-on account compromise.

Security

SCKit ScreenConnect Attack: MSPs Must Hunt Now

Microsoft Threat Intelligence confirmed a campaign abusing ConnectWise ScreenConnect to deploy unauthorized remote-access clients and the SCKit worm on MSP-managed endpoints. The activity can also target technician sessions, making this a high-priority incident for defenders because removing a rogue client alone does not eliminate persistence or invalidate exposed sessions.

Security

AI Agent Supply Chain Attacks Target Coding Tools

Microsoft Security researchers detailed two post-training supply chain attack paths that can backdoor coding and terminal agents: dataset poisoning and model poisoning. The findings matter because compromised agents can misuse tools, execute commands, and access sensitive environments even when prompts appear benign, reinforcing the need for stronger controls around training data, model provenance, and tool permissions.