genai
SECURITY LAB
Learn · Free hands-on series

Learn AI security by breaking it.

A free series that teaches LLM and AI application security the way you actually test it: short concept material, then a live in-browser lab where you run the loop — Attack → Defend → Verify. You exploit a real target, implement the fix, and prove the fix holds. No setup, no API keys, no Docker.

Reading is open, no paywall. Labs are free with a sign-in.
No setupNo API keysNo Docker
Attack → Defend → Verify
01 · ATTACK
Exploit a real target.
02 · DEFEND
Implement the fix.
03 · VERIFY
Prove the fix holds.

Who is this for?

  • Penetration testers and AppSec engineers who already know web application security and need to add AI/LLM components to their testing scope. Applications are shipping with LLM features now, and clients are starting to ask for AI testing.
  • Security engineers and developers defending AI systems — the defensive track of every part walks you through implementing the fix and checking it holds.

What do I need to start?

  • Familiarity with web application security concepts (injection, access control, SSRF-style trust boundaries) is enough.
  • No ML knowledge required. LLMs are treated as components in an application — the security testing mindset transfers directly.

You already know the shape of these bugs

LLM vulnerabilities are the trust-boundary bugs you already hunt, on a new surface. The exploitation instinct transfers — but the defenses don’t, and that’s the part worth learning.

If you can find this……you’ll recognize thisThe catch
SQL injection Prompt injectionSQLi has parameterized queries. Prompt injection has no equivalent — every defense is probabilistic.
Stored XSS Indirect prompt injectionThe attacker poisons content the model reads later — a web page, email, or retrieved doc. Closer to stored XSS than SQLi.
Reflected XSS / template injection Improper output handlingModel output rendered or executed unsanitised → XSS, SQLi, RCE downstream.
SSRF Tool abuse / excessive agencyOver-scoped tools and standing identities turn a prompt into a confused-deputy chain.
IDOR / broken access control Context & canary leakageThe model treated as an authorization boundary that isn't one.

The one-sentence version: prompt injection is to LLM apps what SQLi was to web apps in 2002 — the canonical first bug — except the industry hasn’t found its parameterized queries yet, and may not. That’s why the defensive track here isn’t “use prepared statements”; it’s layered controls you verify under attack.

Why does this exist?

Ten years ago I published a free Android application security series at manifestsecurity.com. At the time, Android security had almost no free material to learn from — web security had plenty. That gap is why I wrote it. It still gets traffic today, and people still message me saying it’s how they learned Android pentesting.

— Aditya Agrawal, Founder GenAI Security Lab

The curriculum

Two paths are live now, and a new part ships every week — each mapped to the OWASP LLM Top 10 (2026), the Agentic AI Top 10, and the MCP Top 10, so you can cite a recognized framework in your reports.

New to the terminology? Browse the AI security glossary

Foundations

AI Security Foundations

Build the map before you specialise. The trust-model shift that makes LLM security different, just enough of how a model works, the attack surface where every OWASP LLM risk lives, and how it all maps to the AppSec you already know.

  1. 1Why LLM security is different
  2. 2How LLMs work, just enough
  3. 3Anatomy of an LLM app: the attack-surface map
  4. 4Prompts: engineering vs injection
  5. 5The LLM threat model, mapped to AppSec
  6. 6Orientation: OWASP, the two tracks, and scope
  7. 7The lab toolkit: how this platform works
Free7 parts, in orderStart the path
OWASP LLM01 · 2026

Prompt Injection

The attack every other LLM exploit rides on. Work the full technique tree — direct, jailbreaks, encoding, smuggling, context, indirect, and tool- and agent-mediated — then the controls that actually hold, each with a live target to try it on.

  1. 1Direct prompt injection
  2. 2Jailbreaks and guardrail evasion
  3. 3Encoding and obfuscation
  4. 4Smuggling and hidden injection
  5. 5Context and example manipulation
  6. 6Indirect prompt injection
  7. 7Tool, memory, and agent-mediated injection
  8. 8Controls that actually hold
Free8 parts, in orderStart the path
No lab setup. Everything runs hosted in your browser — open the part, exploit the target, write the fix, verify. It works the same on a pentest engagement machine as on your laptop.
New part every week — or jump ahead

The next parts are dated below. Every part above is already available in full on the paid learning path — subscribe and work the whole series today.

Get the full path
OWASP LLM02
Sensitive Information Disclosure

Pull the hidden system prompt, secrets sitting in context, and training-data recitation out of the model.

OWASP LLM08
Hidden Context Exposure

Surface context the app never meant to return — RAG (retrieval-augmented generation) leakage and cross-tenant bleed from a shared retrieval pipeline.

OWASP LLM09
Vector & Embedding Weaknesses

Attack the vector store itself — embedding inversion, cross-tenant retrieval, and poisoned indexes.

OWASP LLM10
Improper Output Handling

Model output rendered or executed unsanitised — XSS, SQL injection, and stored second-order injection downstream.

OWASP LLM03
Excessive Agency

Turn an agent's tools against it: over-scoped actions and the confused-deputy chain that makes a prompt trigger a privileged call.

OWASP LLM04
Supply Chain

Compromise what the app trusts before runtime — poisoned models, tampered dependencies, and unvetted third-party components.

OWASP LLM05
Data & Model Poisoning

Corrupt what the model learns or retrieves — poisoned training data, fine-tunes, and RAG sources that steer it.

Cross-cutting
OWASP Agentic AI Top 10

The agent-specific risks: goal manipulation, unsafe tool use, memory poisoning, and multi-agent trust failures.

Cross-cutting
OWASP MCP Top 10

Model Context Protocol risks: malicious tool descriptions, over-permissioned connectors, and consent/onboarding injection.

If you find this series useful, share it — same as the Android series, it only grows by people passing it around.

Suggestions, problems, or a topic you want covered: [email protected]

© 2026 GenAI Security Lab. All rights reserved. You may read, quote, and link to this material with attribution. Copying, republishing, redistribution, resale, or use to train models or build competing products is prohibited without prior written permission.