Skip to content
EDWartens UK

Free course · AI and machine learning

AI Security and the OWASP Top 10 for LLMs

The security course for people who have already built something with an LLM: the attack surface a model adds, all ten 2026 OWASP LLM risks one at a time, indirect injection and zero-click agent attacks, what leaks out through a chatbot, securing agents with identity and least privilege, the OWASP MCP and agentic guidance, red-teaming with MITRE ATLAS, and the NIST-based governance a plant or a GCC can actually enforce.

See the modules

Free to learn. Verifiable certificate £19 when you pass the final exam (optional, VAT included).

Level
Intermediate
Modules
15
Video
14 hours
Price
Free

Outcome

What you will be able to do

Threat-model an AI feature by its parts; name and defend against all ten OWASP LLM risks; trace an indirect injection from a poisoned document to an executed action; write the rule for what may be pasted into a public chatbot; check the provenance of a model, dataset or plugin before it is used; scope an agent's tools, identity and permissions so it cannot act beyond its brief; review a connector before wiring it in; plan a pen test of an LLM feature; and fill the register entry and AI-use policy that make the whole thing auditable.

  • Draw an AI feature as its parts and mark where an outsider's text can enter
  • Name all ten 2026 OWASP LLM risks and sort a real incident into the right one
  • Trace an indirect injection from a poisoned document to an executed action
  • Write the one-page rule on what may and may not be pasted into a public chatbot
  • Scope an agent's tools, identity and permissions to the least it actually needs
  • Plan a pen test of an LLM feature and fill the AI register entry behind it

Software

None required. A free chatbot account is enough for the leakage and grounding exercises.

Hardware

None.

Syllabus

15 modules

Each module has video lessons, notes and a short quiz. The final exam closes the course.

  1. 1Why an LLM is a new attack surface4 lessons
  2. 2The 2026 OWASP Top 10 for LLM applications2 lessons
  3. 3LLM01: prompt injection and its defences4 lessonsLab
  4. 4Indirect injection, promptware and zero-click attacks3 lessonsLab
  5. 5LLM02 and LLM08: what leaks out4 lessonsLab
  6. 6LLM04 and LLM05: supply chain, data and model poisoning4 lessons
  7. 7LLM10 and LLM03: output handling and excessive agency4 lessonsLab
  8. 8LLM09 and LLM07: retrieval weaknesses and misinformation4 lessonsLab
  9. 9LLM06: unbounded consumption and the cost of running AI4 lessons
  10. 10Securing AI agents: architecture and zero trust4 lessons
  11. 11Agent identity, privilege and delegation3 lessons
  12. 12OWASP guidance for agentic applications and MCP2 lessons
  13. 13Testing AI systems: MITRE ATLAS, pen-testing, adversarial AI3 lessons
  14. 14Governance: NIST AI RMF and a rule people can follow4 lessons
  15. 15Final assessment0 lessons

How it works

How this course works

  1. Watch the video lessonsOne module at a time, at your own pace.
  2. Take the module quizA short quiz at the end of each module shows what stuck.
  3. Keep your notesWrite in My notes beside each module and ask the trainer if you are stuck.
  4. Practise in the labOptionalTry a task from the module in the EDWartens Lab, in your browser.
  5. Sit the final examPass it once you have been through the modules.
  6. Certificate, if you want oneOptionalPaid, and only after you pass. Anyone can check it online.

EDWartens Lab

Practise this in the EDWartens Lab

Try the defences from the course on a safe, made-up system in the AI Workbench: injection resistance, approval before action, data kept out and answers grounded in sources.

Works alongside AI Security and the OWASP Top 10 for LLMs. The Lab is ai for industry practice in your browser: no install, no hardware, free to use. Sign in as a learner and it opens with your account.

Open the free lab See the Lab
AI for industry: Build AI models.

Levels to try with this course

Optional. Skip it whenever you like: it never blocks your progress or your certificate.

Why practise in a lab

Technical work is judged on what you can do: wiring, commissioning, finding a fault. Doing a task yourself teaches more than watching it a second time, and a lab lets you try it safely before you meet real equipment. It does not replace working on real hardware.

  • Practise safely, so a mistake costs nothing.
  • The lab checks your work like a site test, so you know when it is right.
  • Free, in your browser, with nothing to install.

Certificate

An optional certificate when you pass

Learning is free. Pass the final exam and you can have a certificate for £19, including VAT, with a number employers can check online. It records that you completed a free online course; it is not a regulated qualification or a CPD certificate.

  • Your name, the course, its study time and your final exam score
  • A unique certificate number and a QR code to its verification page
  • Issued by Wartens Ltd; add it to your LinkedIn profile in one click
  • Pay now or after you pass: it is issued the moment you pass the final

How the certificate works

Sample certificate for AI Security and the OWASP Top 10 for LLMs

Credits

Who made the lessons

The video lessons belong to their creators. EDWartens wrote the study plan, notes, quizzes and the final exam.

  • IBM Technology

    the attack-surface, prompt injection, promptware kill chain, zero-click, shadow AI, corpus poisoning, LLMjacking, agent security, zero trust, MITRE ATLAS and NIST AI RMF explainers

  • Vandana Verma

    the risk-by-risk walkthrough of LLM01 to LLM10, the first look at the OWASP MCP Top 10, and the OWASP Securing Agentic Applications guide

  • Jeff Crume

    the long adversarial AI session on evasion, model inversion, model theft and deepfakes

Questions

About this course

Who is it for?
Engineers who have already put an LLM, a RAG chatbot or an agent near company data, and IT or OT security people who now have AI systems on their asset list. It is the security counterpart to the building courses in the AI track.
What do I need to know first?
You should have seen an LLM application from the inside, so any LLM, RAG or agent course in the track first. Some IT or OT security background helps but is not assumed, and no coding is required. Every task is done on paper against a system you already work with.
Is this the OWASP certification?
There is no OWASP certification for the LLM Top 10. The OWASP Top 10 for LLM Applications is a free published document that anyone may read, and this course teaches you to apply it to a real system. This course is not a security certification of any kind and does not prepare you for one.
Is this a hacking course?
No. It is defensive: threat modelling, controls, agent permissions, governance and incident response. The testing module covers how an authorised test of an LLM feature is scoped, run and written up, not how to attack somebody else's system.
How long is it?
Fourteen modules plus a final, about nine hours of video and roughly fourteen hours in total once the notes, practice tasks and practice questions are counted. A module a day for a fortnight works well.
Is the course really free?
Yes. Every module, practice task and assessment. You create an account so your progress is saved and the assessments can be marked. The certificate is the only paid item, and only if you want it.
What certificate do I get?
An EDWartens Certificate of Completion, issued when you have finished every module and passed the final at 60 per cent, with a verification code anyone can check. It is not a vendor credential and is never described as one.
Who made the video lessons?
The creators named in the Credits block at the foot of this page, on their own YouTube channels. EDWartens did not make the videos and the creators are not affiliated with EDWartens. What EDWartens wrote is the study plan, the notes, the practice tasks and the assessments.

The Professional Module is five live days of Siemens PLC, HMI and SCADA with an instructor, then a resume session and a CPD certificate.

See the Professional Module