My AUGMENTED AI Study Framework for Cybersecurity

5 min read · AI · Study Skills

Watch the full video

Prefer watching? This post is based on my YouTube video.

All formal education is in a crisis right now, with students being lazier than ever using AI for every assignment. Let’s change that, and use AI for augmentation.

Part 1: Recon and narrowing the scope

Identify a target concept, topic, or thing you want to learn. For example: “I want to improve my bug bounty skills.”
Wrap this topic into a “project” or “lab.”
Get direct instructions for the setup, but not the execution of the lab.

Here’s the full prompt I use:

Prompt

I want to improve my ______ skills.

Your task is to act as a cybersecurity mentor and design a practical weekend lab for me.

1. Identify 1–3 high-quality, hands-on platforms or labs (CTFs, fundamentals, or guided labs) that are best for improving this specific skill.
2. Prioritize platforms that are interactive and widely used. Please search the web and find the best platform available for my topic.

For each resource:
- Explain WHY it is relevant (1–2 sentences max)
- Provide exact step-by-step SETUP instructions (account creation, navigation, what to click)
- Tell me exactly where to START (specific room, lab name, or path)

Constraints:
- Do NOT explain the concepts in detail
- Do NOT solve the labs
- Focus only on actionable setup and starting steps

End with a simple execution plan for the weekend (e.g., "Day 1: X, Day 2: Y").

Also include common beginner mistakes or time-wasting paths I should avoid for this skill.

The ideal result

The AI goes out and finds resources such as Bugforge.io, Burp Suite labs, TryHackMe, and even online resources, guides, or cheat sheets. It ideally gives step-by-step instructions for each platform. For example, on TryHackMe:

Navigate to tryhackme.com, create an account, and join the “NahamStore” room. If you get stuck, there are many walkthroughs online, including one from the creator, NahamSec.

Part 2: Execute and iterate

The core idea: you only learn cyber by doing and failing fast. Friction is good. Here’s the loop:

Execute

Do the thing. Write the exploit, run the nmap scan, trace the packet, respond to the incident.

Get stuck

Move forward until something breaks or you don’t know what to do next. Then ask yourself: what happened, and why did I get stuck?

Ask AI specific failure questions

Detail the mismatch between the expected and actual system behavior.

Prompt

Here's what I tried:
[exact action]

Here's what happened:
[exact result]

What are 2–3 likely reasons for this, and what should I try next? Explain simply why this is a good next step.

No full solutions. Only next steps and possibilities.

Retry

Take ONE suggestion, apply it, and run the test again.

Understand

Explain it in your notes in 1–2 sentences: “This worked because ___.”If you can’t explain it like you would to a 5 year old, you don’t truly know it yet. This is also known as the Feynman Technique.To reinforce your learning, validate with the LLM of your choice. Put your reasoning in and ask, “Is this reasoning correct?”

Wrapping up

AI should give you direction if you’re lost, but not the entire map. That is something that you can figure out yourself. The whole idea of this framework is a loop:

You execute
You fail
AI refines your thinking
You execute again

Repeating this process over and over is what learning looks like.

⌘I

Watch the full video

​Part 1: Recon and narrowing the scope

​The ideal result

​Part 2: Execute and iterate

​Wrapping up

Part 1: Recon and narrowing the scope

The ideal result

Part 2: Execute and iterate

Wrapping up