AI is racing into production environments faster than security teams can keep up—and Oligo Security, the runtime security company, is stepping directly into that gap. Today the company unveiled a sweeping set of new capabilities purpose-built to secure the full spectrum of AI technologies in production, from LLMs to autonomous agents, addressing what many experts now call the single largest blind spot in modern cybersecurity.
“AI is moving into production faster than it can be secured, forcing businesses to take greater risks in the name of speedy innovation,” said Nadav Czerninski, co-founder and CEO of Oligo Security. “Runtime is the only source of truth for how systems actually behave—especially in AI, where agents can execute code on the fly and models live inside the production stack.”
Oligo’s new platform modules—AI-SPM (AI Security Posture Management) and AI-DR (AI Detection & Response)—extend the company’s proven runtime approach to the emerging challenges of operational AI.
A Proven Runtime Approach, Now Applied to AI
Oligo’s runtime sensor has already demonstrated its ability to detect and prevent real-world AI attacks. Using its telemetry, Oligo researchers:
- Identified ShadowRay, the first known attack targeting AI workloads in the wild (March 2024)
- Uncovered ShadowRay 2.0 (November 2025), the first self-replicating AI-driven botnet
- Surfaced vulnerabilities in high-risk AI systems from Anthropic, Meta, NVIDIA, Ollama, and PyTorch
These discoveries give Oligo a unique understanding of how AI workloads behave under real-world conditions—a foundation that now powers its expanded platform.
AI-SPM: Real-Time AI Security Posture Management
As organizations deploy more LLMs, agents, and external AI services, most have no accurate inventory of what’s running in production. AI-SPM closes that gap by providing a complete, continuously updated view of every AI component in the environment.
Key Capabilities
- Runtime AI Inventory
Live detection of all models, agents, SDKs, frameworks, and third-party AI services. - Continuous AI-BOM
An always-current bill of materials for AI components and dependencies. - Runtime-Based Risk Detection
Flags unsafe, untrusted, or vulnerable models based on how they actually behave—not assumptions. - Compliance & Governance
AI usage mapped to emerging regulatory frameworks with audit-ready reports.
Oligo positions AI-SPM as the first practical solution for governing AI at scale, rooted in ground-truth runtime insights rather than static documentation.
AI-DR: Real-Time Detection and Response for AI Threats
Modern AI introduces entirely new operational risks—from jailbreaks to rogue agent behavior—and these risks often manifest only at runtime. AI-DR gives security teams the visibility, policy enforcement, and automated response mechanisms they need to contain threats before damage occurs.
Key Capabilities
- Model Protection
Defense against prompt injection, jailbreaks, and unsafe model outputs. - Runtime Agent Protection
Monitoring and control over agent tool usage, code execution, and API calls. - Automated Response
Real-time blocking, containment, and integration with SOC workflows and existing SIEM/EDR systems.
Together, AI-SPM and AI-DR establish the industry’s first unified runtime protection model for both traditional workloads and next-generation AI systems.
Customer Perspective: Production AI Requires Production-Grade Security
“AI is at the core of our business, making visibility and control in production a non-negotiable,” said Robert Kugler, Head of Security, IT, and Compliance at Cresta. “With runtime as our source of truth, Oligo has drastically improved our ability to understand and mitigate risk—while also protecting us against unknown threats through their anomaly detection.”
Why This Matters: Closing the AI Security Blind Spot
AI’s rapid move into production has outpaced traditional security tooling. Models and agents now:
- Generate code dynamically
- Access internal APIs and databases
- Run autonomously within production environments
- Interact with sensitive customer and business data
These behaviors create new attack surfaces that can only be observed—and secured—at runtime.
With these new modules, Oligo is positioning itself as the first company capable of providing full-stack runtime security for AI, solving a challenge most organizations are only beginning to confront.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
