RAVINDRA ANNAM
AI & Cyber Security Architect | Agentic AI Security Researcher
Securing autonomous AI systems through runtime authorization, authority-aware threat modeling, and security controls for AI agents, tools, memory, and delegation.
Agentic AI Security • AI Threat Modeling • Runtime Security • Runtime Authorization • LLM Security
20+ Years
Cybersecurity Experience
AI Security
Research & Architecture
Published
Research & Industry Author
Professional Service
Peer Reviewer & Industry Judge
Profiles
Google Scholar | ORCID | ResearchGate | Zenodo
About Ravindra Annam
Ravindra Annam is an AI & Cyber Security Architect with over 20 years of experience specializing in AI Security, Agentic AI Security, Application Security, Cloud Security, Threat Modeling, DevSecOps, and Enterprise Product Security Architecture.
His work focuses on securing AI agents, LLM-powered applications, MCP ecosystems, RAG systems, APIs, autonomous workflows, and cloud-native enterprise platforms through secure-by-design and runtime security approaches.
Ravindra contributes to the cybersecurity community through technical publications, conference and university speaking, peer review, industry judging, AI security research, and professional security initiatives.
Areas of Expertise
AI Security • Agentic AI Security • LLM Security • Runtime AI Security • Application Security • Threat Modeling • Cloud Security • MCP Security • RAG Security • DevSecOps • AI Governance
The AI Runtime Threat Matrix: A Threat Taxonomy and Control Framework for Autonomous AI Agent Systems
Author: Ravindra Annam
Publication type: Technical Report
Version: 1.0
Publication date: August 23, 2026
Publisher/Repository: Zenodo
DOI: 10.5281/zenodo.22063227
Abstract
Autonomous AI agents introduce security risks that extend beyond traditional application and model-centric threat models. Agents can dynamically select tools, invoke APIs, access memory, interact with external systems, delegate tasks, and perform privileged actions across multi-step workflows. These capabilities create runtime attack paths in which untrusted influence can propagate into consequential actions.
This technical report introduces the AI Runtime Threat Matrix, a structured threat taxonomy and control framework for analyzing security risks across autonomous AI agent systems. The framework organizes runtime threats across identity, authorization and permissions, prompts and context, memory, tools and APIs, data flows, external systems, multi-agent interactions, autonomous execution, human oversight, monitoring, and governance.
Keywords: Agentic AI; AI Security; Autonomous AI Agents; Runtime Security; Threat Modeling; Runtime Authorization; Authority Escalation; Prompt Injection; Memory Poisoning; Tool Security; MCP Security; Multi-Agent Systems; AI Agent Security; Cybersecurity
Annam, R. (2026). The AI Runtime Threat Matrix: A Threat Taxonomy and Control Framework for Autonomous AI Agent Systems (Version 1.0). Zenodo. https://doi.org/10.5281/zenodo.22063227
