Udemy - Prompt Injection and Jailbreaking - Offensive LLM Red-Tea...

Prompt Injection & Jailbreaking: Offensive LLM Red-Teaming

https://WebToolTip.com

Published 7/2026
Created by Dr. Amar Massoud
MP4 | Video: h264, 1920x1080 | Audio: AAC, 44.1 KHz, 2 Ch
Level: Intermediate | Genre: eLearning | Language: English | Duration: 62 Lectures ( 5h 31m ) | Size: 2.6 GB

Hands-on offensive LLM security: jailbreaks, indirect injection, agent exploitation, and defenses in a sandbox.

What you'll learn
⚡ Explain at the architecture level why LLMs cannot separate instructions from data
⚡ Craft and chain direct jailbreaks that bypass a model's refusal layer
⚡ Defeat keyword and moderation filters using encoding and Unicode evasion
⚡ Plant indirect prompt injections in RAG documents, emails, and web content
⚡ Hijack agent tool calls to exfiltrate secrets and abuse real functions
⚡ Deliver payloads through image, OCR, and other multimodal channels
⚡ Run automated red-team loops and adversarial-suffix attacks at scale
⚡ Map guardrails, bypass hardened defenses, and write a professional findings report

Requirements
❗ Working knowledge of LLM and API basics (system vs. user prompts, tokens, embeddings)
❗ Comfortable reading and running Python scripts
❗ Basic web and application security concepts (trust boundaries, SSRF, injection)
❗ A machine capable of running a local open-weight model for the labs
❗ Willingness to operate strictly within authorized, sandboxed testing