Nov 19, 2026
HackTrack With Mentors: Take On AI-300
If OSAI is on your list of certifications to earn in 2027, make room for HackTrack guided sessions every Saturday!
AI-300 is our newest course, and month after month, it’s among our most enrolled. There’s a clear appetite for learning how to attack AI-enabled systems. Keeping up with that learning alongside work and everything else takes commitment.
HackTrack with Mentors is a 12-session, mentor-led learning program that helps AI-300 learners strengthen their AI red teaming skills through guided practice, practical demonstrations, and community support.
Starting November 28, join OffSec Student Mentors and fellow learners each Saturday to work through the material and stay accountable on your learning journey.
If OSAI is on your list of certifications to earn in 2027, make room for these guided sessions on your Saturdays.
Learning how an exploit works is one thing but recognizing where to use it, interpreting an unexpected result, and deciding what to try next takes practice.
HackTrack gives you an opportunity to follow that process with OffSec Student Mentors. Through practical demonstrations and guided exercises, mentors will explain their approach to assessing AI-enabled systems, including the troubleshooting that helps an assessment move forward.
Use the sessions to:
- Work through difficult concepts: Bring questions from your studies and get guidance on the parts you haven’t quite figured out.
- Understand the decisions behind an attack: Follow how mentors investigate a target and choose their next step.
- Build beyond AI-300: Explore additional skills and custom content, including open-source models and hackbots.
Between sessions, bring what you’ve learned back to your AI-300 studies. The next Saturday gives you an opportunity to return with questions and build on that work, with mentors available to help you through the parts that are slowing your progress.
AI red teaming involves more than testing what a chatbot will say. AI-300 develops the offensive skills to assess these systems. HackTrack supports that learning by helping you connect individual techniques to a broader testing methodology.
Across the 12 sessions, you will be examining areas such as:
- Agent exploitation: Assess attack surfaces and investigate weaknesses in agent workflows.
- RAG pipelines: Work through retrieval hijacking, ingestion poisoning, and knowledge base leakage.
- MCP and tool surfaces: Examine how integrations expand what an AI-enabled application can access.
- Infrastructure and supply chains: Investigate weaknesses in the systems and dependencies supporting AI deployments.
Attending the full series helps you understand how those techniques relate to one another. In the final two sessions, you’ll bring them together and reinforce a repeatable approach to testing AI-enabled targets.
A busy week can easily spiral into two weeks away from a course. Committing to all 12 Saturdays gives your studies a regular place in your calendar and a community to return to.
You’ll learn alongside others working through AI-300, hear questions you might not have thought to ask, and have opportunities to work through difficult concepts together. As the sessions progress, you can revisit earlier lessons with a better understanding of how they fit into an assessment.
Whether you’re just beginning the course or preparing for OSAI, that sustained practice helps you get more from the material.
The program also includes a Certificate of Attendance.
| Week | Session | Highlights | Methodology Focus |
| 1. Nov 28 | Introduction to AI-300 | LLMs, chatbots, RAG, and agents; AI hosting and model benchmarking; agent setup; responsible AI-assisted testing | Build a foundation in AI red teaming, understand the AI-300 learning environment, and explore how to use AI tools responsibly during security assessments. |
| 2. Dec 5 | AI Recon and Recon with AI | Passive reconnaissance; active reconnaissance; testing AI chatbots and agents; AI-assisted information gathering | Develop a structured reconnaissance methodology for identifying AI attack surfaces and using AI to support security testing. |
| 3. Dec 12 | Attacking AI Agents | AI agent attack surfaces; agent exploitation techniques; guided exploration of agent security | Strengthen your understanding of how AI agents can be assessed and where exploitable weaknesses may emerge. |
| 4. Dec 19 | Multi-Agent Systems & A2A | Multi-agent architectures; agent discovery; trust boundaries; Agent Cards; workflow manipulation; indirect prompt injection | Learn how to map communication and trust relationships between agents and identify opportunities to exploit multi-agent workflows. |
| 5. Jan 9 | Exploiting RAG Pipelines and Embeddings | RAG architecture; knowledge base leakage; ingestion poisoning; embedding collisions; retrieval hijacking; defense evasion | Explore how retrieval and embedding processes introduce new attack surfaces, and develop a methodology for testing RAG-enabled applications. |
| 6. Jan 16 | Attacking Model Context Protocol and Tool Surfaces | Model Context Protocol (MCP); AI tool surfaces; security testing of tool-connected AI systems | Understand how AI integrations and connected tools expand the attack surface of AI-enabled applications. |
| 7. Jan 23 | Open-Source Models and Hackbots | Open-source AI models; hackbots; custom learning content beyond AI-300 | Expand your understanding of open-source AI tooling and its role in offensive security workflows. |
| 8. Jan 30 | Supply Chain Attacks on AI/ML Systems | AI/ML supply chains; supply chain attack surfaces; security risks across AI system dependencies | Develop an approach to identifying and evaluating security weaknesses throughout the AI/ML supply chain. |
| 9. Feb 6 | AI Infrastructure and Deployment Exploits | AI infrastructure; deployment environments; infrastructure attack surfaces | Explore how the systems supporting AI models and applications introduce additional exploitation opportunities. |
| 10. Feb 13 | Threat Modeling for AI-Enabled Targets | AI threat modeling; attack surface identification; evaluating potential attack paths | Strengthen your ability to systematically identify threats, assess exposure, and prioritize testing across AI-enabled environments. |
| 11. Feb 20 | Assembling the Pieces – Part 1 | Applying concepts from previous sessions; connecting AI attack techniques; structured assessment methodology | Begin bringing individual techniques together into a broader approach to AI red teaming. |
| 12. Feb 27 | Assembling the Pieces – Part 2 | Reinforcing AI red teaming concepts; connecting attack paths; applying assessment methodology | Bring together the concepts explored throughout HackTrack and reinforce a repeatable approach to testing AI-enabled systems. |
HackTrack with Mentors is free for AI-300 learners with active lab access, for learners that are studying individually or through your organization.
- Starts: Saturday, November 28, 2026
- Time: 16:00 GMT
- Duration: 12 Saturday sessions, ending February 27, 2027
- Holiday break: No sessions on December 26 or January 2
Bring all of the questions you’ve been saving and left unanswered. See you at the first session!
Get Learn One for 20% off, and start HackTrack for Session 1!