Course Objective
What this programme helps you achieve
A hands-on 2 days programme that builds practical, job-ready capability in this area. Delivered through guided walkthroughs, worked examples and reusable checklists you can apply immediately.
UnderstandBuild a clear, structured understanding of the core concepts, terminology and drivers behind this domain.
ApplyTranslate the framework into practical actions, controls and working methods you can use on real engagements.
AssessEvaluate current-state maturity, identify gaps and prioritise the improvements that reduce the most risk.
DeliverProduce the artefacts, evidence and reporting expected of a competent practitioner in this area.
Market Need
Why this matters now
Prompt Injection, RAG Security and Agentic AI Security Testing addresses a capability that organisations increasingly need but often lack in a structured, defensible form. Teams frequently rely on ad-hoc knowledge, scattered documents or vendor claims instead of a repeatable method.
This programme closes that gap. It gives participants a shared vocabulary, a practical method and a set of reusable artefacts so the organisation can operate this domain consistently rather than reactively.
◆
Framework alignment: OWASP LLM Top 10 and agentic AI security testing practices.
What makes this programme useful
01
Practical Focus
Every module is grounded in real scenarios, walkthroughs and reusable checklists rather than theory alone.
02
Career Relevant
Skills map directly to in-demand roles and to how organisations actually operate this domain.
03
Structured Path
A logical progression from fundamentals to applied practice, with a clear takeaway toolkit.
Key Learning Areas
What participants will learn
Direct and indirect prompt injection techniques
Jailbreaks, system prompt leakage and bypass patterns
Retrieval-Augmented Generation (RAG) architecture risks
Poisoning and untrusted content in RAG pipelines
Agentic AI, tool use and excessive agency risks
Securing tool calling, function execution and permissions
Testing methodology for LLM and agentic systems
Mitigations: input/output filtering, isolation and monitoring
2 Days Agenda
Structured learning flow
Day 1
Day 1 — Prompt Injection & RAG Attacks
Direct and indirect prompt injection techniques; Jailbreaks, system prompt leakage and bypass patterns; Retrieval-Augmented Generation (RAG) architecture risks; Poisoning and untrusted content in RAG pipelines.
Day 2
Day 2 — Agentic Testing & Defences
Agentic AI, tool use and excessive agency risks; Securing tool calling, function execution and permissions; Testing methodology for LLM and agentic systems; Mitigations: input/output filtering, isolation and monitoring.
Audience & Entry Level
Who should attend?
This programme is designed for professionals who work with, or are moving into, this area of cybersecurity, governance or technology.
AI Governance LeadRisk & ComplianceData Science / ML TeamSecurity ArchitectProduct & EngineeringCISO / CTOIT AuditorLegal & Privacy
Recommended prerequisite
A general awareness of information security or IT concepts is helpful. No specific certification is required to attend.
Learning Outcome
What participants should be able to do after the course
- Confidently work with direct and indirect prompt injection techniques.
- Confidently work with jailbreaks, system prompt leakage and bypass patterns.
- Confidently work with retrieval-augmented generation (rag) architecture risks.
- Confidently work with poisoning and untrusted content in rag pipelines.
- Confidently work with agentic ai, tool use and excessive agency risks.