Module 01 — AI, Machine Learning & Security Fundamentals

Establish a clear technical foundation in AI, machine learning, generative AI and the security challenges introduced by modern AI systems.

Core Topics

• Introduction to Artificial Intelligence and AI Security
• AI vs. Machine Learning vs. Deep Learning vs. Generative AI
• Traditional Programming vs. Machine Learning
• How Machine Learning Systems Learn
• Neural Networks: Layers, Weights and Learning Concepts
• AI Models, Datasets, Training, Prompts and Context
• Introduction to Large Language Models (LLMs)
• How Modern AI Applications Work
• AI Security vs. Traditional Cybersecurity
• AI Security vs. Using AI for Cybersecurity
• AI Threat Landscape and AI-Specific Attack Categories
• AI Attack Surface and Security Challenges
• Responsible and Secure AI Usage

Practical Application

AI Attack Surface Identification

• Analyse a controlled AI application and identify its components, trust boundaries and potential exposure.
• Produce an AI attack-surface map.

Module 02 — AI Application Architecture & Reconnaissance

Understand modern AI application architecture and identify exposed components before security testing.

Core Topics

• Anatomy of an AI Application
• LLM-Powered Applications and AI APIs
• Model Endpoints and System Prompts
• Context Windows and Conversation Memory
• RAG Architecture and Knowledge Bases
• Vector Databases — Security Perspective
• AI Agents and External Tools
• Third-Party Integrations and APIs
• AI Application Trust Boundaries
• AI Reconnaissance Methodology
• Identifying Exposed AI Components
• AI Application Fingerprinting
• AI Attack-Surface Mapping

Practical Application

AI Reconnaissance Lab

• Perform reconnaissance against a controlled AI application.
• Document models, endpoints, prompts, data sources, integrations and trust boundaries.

Module 03 — AI Models, Frontier Models & Model Security

Develop a security-oriented understanding of general-purpose, reasoning, multimodal, local and frontier models.

Core Topics

• General-Purpose AI Models and Large Language Models
• Reasoning, Coding and Multimodal Models
• Small Language Models and Local Models
• Open-Source vs. Proprietary Models
• Cloud Models vs. Local Models
• Frontier Models and Emerging Capabilities
• Model Capabilities, Limitations and Evaluation
• Tokens, Context Windows and Model Behaviour
• Tool-Using and Agentic Capabilities
• Model Selection for Security Research
• Model Security Considerations
• Security Implications of Model Deployment Choices

Practical Application

AI Model Security Comparison

• Compare model types from a security-testing perspective.
• Evaluate capabilities, limitations and deployment-related risks.

Module 04 — AI Threat Modeling & Risk Assessment

Systematically identify threats, trust boundaries, attack paths and security risks across AI applications.

Core Topics

• Introduction to AI Threat Modeling
• AI vs. Traditional Threat Modeling
• Assets, Data Flows and Trust Boundaries
• AI Attack Surfaces and Threat Actors
• Attack Paths and AI Abuse Cases
• Security Objectives and Threat Identification
• Risk Assessment and Impact Analysis
• Authentication and Authorization
• RAG and AI Agent Threat Modeling
• AI Data Security
• STRIDE for AI Applications
• OWASP Security Approach
• MITRE ATLAS Introduction
• AI Security Controls and Risk Treatment
• Security Findings, Risk Ratings and Reporting

Practical Application

AI Threat Model Assessment

• Create a complete threat model for a simulated enterprise AI application.

• Identify assets, threats, attack paths, impact and controls.

Module 05 — LLM Security & OWASP Top 10

Perform structured security testing of LLM-powered applications using recognized vulnerability categories.

Core Topics

• LLM Security Architecture
• LLM Attack Surface
• Prompt and Input Security
• Input and Output Validation
• Insecure Output Handling
• Sensitive Information Disclosure
• System Prompt Exposure
• Improper Handling of Untrusted Input
• Excessive Agency
• Model Denial of Service
• Supply-Chain Vulnerabilities
• Vector and Embedding Security
• Overreliance on LLM Output
• Prompt Injection Chains
• Security Misconfiguration
• Automated LLM Security Testing
• Safe and Authorized LLM Security Assessment

Practical Application

LLM Security Assessment Lab

• Assess a controlled LLM application against common security weaknesses.
• Document evidence, impact, risk and remediation.

Module 06 — Prompt Injection & Instruction Attacks

Assess how untrusted instructions influence AI systems through direct and indirect prompt-injection scenarios.

Core Topics

• Prompt Injection Fundamentals
• Direct Prompt Injection
• Indirect Prompt Injection
• Injection Through External Content
• System Prompt Manipulation
• Context Manipulation
• Instruction Conflicts and Instruction Hierarchy
• Prompt Boundary and Trust Issues
• Prompt-Based Data Exfiltration
• Document-Based Prompt Injection
• Webpage-Based Prompt Injection
• Multi-Turn and Multi-Step Injection
• Prompt Injection Against AI Assistants
• Prompt Injection Against RAG Applications
• Prompt Injection Against AI Agents
• Detection and Defensive Prompt Design
• Input Isolation, Context Separation and Output Validation

Practical Application

Prompt Injection Lab Series

• Direct Prompt Injection
• System Prompt Security
• Indirect Prompt Injection
• Document-Based Injection
• Multi-Step Prompt Attack

Module 07 — Jailbreaking & AI Guardrail Security

Evaluate AI safety controls and understand how security researchers assess guardrails and filters in controlled environments.

Core Topics

• Jailbreaking Fundamentals
• Jailbreak vs. Prompt Injection
• AI Safety Mechanisms and Guardrails
• Safety Filters and Instruction Hierarchy
• Adversarial Prompting Concepts
• Context Manipulation
• Role-Based and Multi-Turn Attacks
• Guardrail Weaknesses and Security Testing
• Model Behaviour Analysis
• Jailbreak Detection
• Input and Output Controls
• Defensive Guardrails
• Human Oversight and Safe AI Security Research

Practical Application

AI Guardrail Security Challenge

• Test a controlled AI system against defined adversarial scenarios.
• Document behaviour, impact, detection, mitigation and retest results.

Module 08 — RAG Security & Knowledge Base Attacks

Assess Retrieval-Augmented Generation systems, knowledge bases and retrieval pipelines that connect private data to LLMs.

Core Topics

• RAG Fundamentals and Architecture
• Documents, Embeddings, Vector Databases, Retrieval and LLMs
• RAG Attack Surface
• Knowledge Base Security
• Vector Database Security
• Retrieval Manipulation
• Malicious Documents
• Prompt Injection Through RAG
• Sensitive Information Disclosure
• Unauthorized Document Retrieval
• Cross-User Data Exposure
• Access-Control Failures
• Untrusted Knowledge Sources
• RAG Trust Boundaries
• Document Integrity and Source Verification
• Retrieval Validation and Security Controls
• Remediation and Retesting

Practical Application

RAG Security Lab

• Assess unauthorized retrieval, information disclosure, malicious content and retrieval manipulation.
• Validate access controls and document remediation.

Module 09 — AI Data Poisoning & Model Integrity

Understand how manipulated data can influence AI behaviour and how to assess data integrity and poisoning risks.

Core Topics

• Data Poisoning Fundamentals
• Training-Data Poisoning
• Fine-Tuning Data Poisoning
• RAG Data Poisoning
• Knowledge Base Manipulation
• Malicious Documents
• Backdoored Datasets
• Dataset Integrity and Data Provenance
• Data Validation and Integrity Verification
• Detecting Poisoned Data
• Model Integrity Considerations
• Secure Data Pipelines
• Monitoring for Data Manipulation
• Data-Poisoning Incident Response

Practical Application

RAG Data Poisoning Challenge

• Analyse a controlled poisoned knowledge base.
• Identify the poisoning vector, assess impact and recommend remediation.

Module 10 — AI Agent & Tool Security

Assess the security implications of agentic AI systems that can plan, access tools, interact with APIs and perform actions.

Core Topics

• Introduction to AI Agents
• LLM vs. Chatbot vs. Assistant vs. Agent
• Agent Architecture
• Planning, Reasoning, Memory and Actions
• Tools and External Integrations
• APIs and Agent Permissions
• Agent Trust Boundaries
• Agent Attack Surface
• Tool Abuse and Excessive Agency
• Privilege-Boundary Failures
• Agent Hijacking
• Indirect Prompt Injection Against Agents
• Multi-Step Attack Chains
• Unauthorized Actions and Sensitive Data Access
• API Security Considerations
• Human-in-the-Loop Security
• Agent Security Controls

Practical Application

AI Agent Security Assessment

• Assess a controlled AI agent with tool/API access.
• Identify unsafe behaviour, excessive permissions and injection paths.

Module 11 — AI Supply Chain Security

Examine risks introduced by models, datasets, dependencies, tools and third-party components.

Core Topics

• AI Supply Chain Fundamentals
• Models and Model Repositories
• Datasets and Data Sources
• Libraries, Frameworks and Dependencies
• Plugins, Tools and AI APIs
• Third-Party AI Components
• Open-Source AI Security Risks
• Malicious Model Artefacts
• Model Tampering and Backdoors
• Compromised Components
• Dependency Risks
• Software and Model Integrity
• Component Verification
• Secure AI Deployment
• AI Supply Chain Risk Assessment

Practical Application

AI Supply Chain Security Assessment

• Analyse a simulated AI project for untrusted or compromised components.
• Recommend controls for verification and secure deployment.

Module 12 — AI Forensics & Incident Response

Investigate AI security incidents, preserve evidence and reconstruct attack activity across LLM, RAG and agent environments.

Core Topics

• AI Security Incident Fundamentals
• AI Attack Indicators
• Prompt Injection Evidence
• Suspicious Model Behaviour
• RAG Compromise Indicators
• Data Poisoning Indicators
• AI Logs and Conversation History
• API and Tool Activity
• Evidence Collection and Preservation
• AI Incident Investigation
• Timeline Reconstruction
• Root Cause Analysis
• Containment, Eradication and Recovery
• AI Incident Response
• Security Reporting

Practical Application

AI Security Incident Response Case

• Investigate a simulated enterprise AI compromise.
• Reconstruct the timeline, identify root cause and prepare an incident report.

Module 13 — AI Penetration Testing & Red Teaming

Bring the programme together through an end-to-end offensive security methodology for AI applications and AI-enabled environments.

Core Topics

• AI Penetration Testing vs. AI Red Teaming
• Adversarial Mindset
• Threat-Actor Simulation
• Attack Objectives and Campaign Design
• AI Reconnaissance and Attack-Surface Discovery
• Prompt and Attack Mutation
• Adversarial Test Cases
• Multi-Turn Attack Campaigns
• LLM Security Testing
• RAG Attack Scenarios
• AI Agent Attack Scenarios
• Tool-Abuse Testing
• Security-Control Validation
• Measuring Attack Success
• Attack Impact Analysis
• Mapping Attacks to Security Risks
• Evidence Collection
• AI Red Team Reporting

Practical Application

AI Red Team Assessment

• Plan and execute a controlled AI security assessment.
• Identify vulnerabilities across LLM, prompt, RAG, agents and tools.
• Validate impact and produce professional findings.

Module 14 — AI Security Frameworks, Standards & Practical Environments

Learn to communicate findings using recognized frameworks and conduct assessments in isolated environments.

Core Topics

• OWASP Top 10 for LLM Applications
• OWASP Generative AI Security Guidance
• MITRE ATLAS
• NIST AI Risk Management Framework
• AI Security Risk Categories
• Mapping Findings to Recognized Security Categories
• Risk Classification and Prioritization
• Security Assessment Methodology
• Local and API-Based AI Models
• Docker-Based Vulnerable Applications
• Vulnerable LLM Applications
• RAG Security Lab Environment
• AI Agent Security Lab Environment
• Safe and Isolated AI Security Testing

Practical Application

Framework Mapping Exercise

• Map findings to OWASP, MITRE ATLAS and NIST-aligned categories.
• Prepare professional risk and remediation statements.

Module 15 — AI Security CTF, Bug Bounty & Security Research

Convert technical knowledge into practical security research capability through controlled challenges, responsible disclosure and a final assessment.

Core Topics

• AI Security CTF Fundamentals
• LLM Security Challenges
• Vulnerable AI Applications
• AI Security Research Methodology
• Vulnerability Discovery and Validation
• Proof-of-Concept Development
• Security Documentation
• Authorized AI Bug Bounty Programs
• Responsible Disclosure
• AI Security Research Communities
• Impact Demonstration
• Responsible AI Security Research

Practical Application

Hackinetix AI Security CTF

• Solve controlled challenges covering LLM security, prompt injection, jailbreaks, RAG, data poisoning, AI agents and supply-chain risks.
• Document findings using a professional security-research format.

Final Capstone Project — Enterprise AI Security Assessment

Learners perform a complete, controlled security assessment of a simulated enterprise AI platform containing an LLM, RAG pipeline, knowledge base, AI agent, external tools, APIs, authentication and security controls.

Assessment Lifecycle

• Reconnaissance
• Attack-Surface Mapping
• Threat Modeling
• Vulnerability Discovery
• Controlled Exploitation
• Impact Validation
• Security-Control Testing
• Remediation Recommendations
• Retesting
• Final Reporting

Final Deliverables

• AI Attack Surface Map
• Threat Model
• Vulnerability Assessment
• Proof of Concept / Evidence
• Risk Ratings
• Mitigation Recommendations
• Retest Results
• Executive Summary
• Technical AI Security Assessment Report

Learning Outcomes

• Understand modern AI, generative AI and LLM architectures.
• Identify AI-specific attack surfaces and security risks.
• Perform AI reconnaissance and attack-surface mapping.
• Threat-model AI systems and assess risk.
• Assess LLM applications against common security weaknesses.
• Evaluate direct and indirect prompt injection.
• Assess AI guardrails and jailbreak resistance in authorized environments.
• Evaluate RAG and knowledge-base security.
• Identify data-poisoning and model-integrity risks.
• Assess AI agents, tools, APIs and excessive-agency risks.
• Evaluate AI supply-chain security.
• Investigate AI security incidents and preserve evidence.
• Conduct structured AI penetration testing and red-team assessments.
• Map findings to recognized AI security frameworks.
• Prepare professional AI security assessment and research reports.

Tools, Frameworks & Technologies

• AI/LLM Platforms: ChatGPT, Gemini, Claude and selected open-source/local LLM environments.
• Security Tools: Burp Suite, API testing tools and browser developer tools.
• Frameworks: OWASP Top 10 for LLM Applications, OWASP Generative AI Security Guidance, MITRE ATLAS and NIST AI RMF.
• Lab Technologies: RAG applications, vector databases, AI APIs, AI agents and intentionally vulnerable AI applications.
• Containerized Environments: Docker-based isolated security labs.