Course Objective & Scope
Traditional application threat modeling alone cannot capture the risks of LLM systems where data and instructions travel in the same channel, decisions are probabilistic and tools are invoked autonomously. This course extends STRIDE with MITRE ATLAS, the OWASP Top 10 for LLM Applications and OWASP Agentic AI taxonomies to systematically uncover novel attack surfaces on data flow diagrams.
Working on a case adapted to their organisation, participants produce a complete AI threat model and prioritised mitigation roadmap covering ingestion pipelines, vector databases, model serving, memory and MCP/tool integrations.
Who Should Attend?
- Security architects and application security (AppSec) engineers
- Software and platform engineers designing AI, RAG and agent systems
- Technical leads participating in risk assessments and secure design reviews
- Prerequisite: familiarity with system architecture and core LLM/RAG components
Key Takeaways
- Draw data flow and trust boundary diagrams incorporating LLM, RAG and agent components
- Map MITRE ATLAS and OWASP LLM/Agentic risks to architectural components
- Construct attack trees across data, model, context window and tool layers
- Score risks by impact, exploitability and action reversibility
- Produce a reusable enterprise AI threat model template for design reviews
Curriculum
From Classical Threat Modeling to AI Threat Modeling
- Differences between deterministic software and probabilistic, autonomous systems
- Integrating STRIDE, MITRE ATLAS, OWASP LLM Top 10 and NIST AI RMF
Architectural Decomposition and Trust Boundaries
- Classifying system prompts, user inputs, RAG sources, memory and tool outputs
- Drawing trust boundaries across ingestion, embedding, inference and action layers
Attack Trees for RAG and Agent Systems
- Modeling indirect injection, knowledge base poisoning and exfiltration chains
- Privilege escalation and goal hijacking scenarios in multi-agent communication
Risk Scoring and Mitigation Design
- Risk assessment matrix factoring in autonomy level and action reversibility
- Risk reduction through architectural controls, guardrails and human approval gates
Hands-On AI Threat Model Workshop
- End-to-end threat modeling on an enterprise AI use case
- Deriving security requirements and a review checklist for design sign-off
Other Courses in This Category
-
Teaches how to plan and execute scenario-driven red team campaigns against LLM, RAG and agent systems using PyRIT, garak and promptfoo, build multi-turn exploit chains and deliver reproducible findings.
- Live Course
- 5 modules
-
Teaches how to design and operate low-latency guardrails across input, retrieval, tool calling and output stages using NeMo Guardrails, Guardrails AI, Llama Guard and deterministic policy engines.
- Live Course
- 5 modules
-
Teaches how to build automated security evaluation suites for LLM and agent applications using promptfoo, garak and Inspect AI, enforce regression gates in CI/CD pipelines and continuously test model and prompt changes.
- Live Course
- 5 modules
-
Teaches how to design enterprise LLM, RAG and agent platforms on Zero Trust principles using a central AI gateway, network and tenant isolation, KMS key management, secure MCP integration and auditable telemetry.
- Live Course
- 5 modules
-
Comprehensive training on PII masking, reversible pseudonymisation, anonymisation and data residency architectures to adopt LLMs without exposing personal or confidential corporate data.
- Live Course
- 5 modules
-
Hands-on training on building a low-latency PII detection, masking and de-pseudonymisation gateway in front of LLM APIs using Microsoft Presidio, GLiNER and Turkish rule/NER recognisers.
- Live Course
- 5 modules
Get in touch
