A comprehensive AI security course

Learn to Hack and
Defend AI Agents

The most comprehensive AI + Cybersecurity course. 12h+ of hands-on content: from LLM fundamentals to prompt injection, agent takeover, and automation of offensive/defensive routines with AI.

While others teach how to use AI, we teach how to attack and defend AI agents with real labs and professional tools.

Think
Secure
Build
Secure Your Spot View Content Tsushin Framework
12h+
Video Lessons
10
Modules
160+
Slides
30+
Labs Hands-on
~8h
Lab Practice
15+
Tools
scroll
MSRC Leaderboard 2022
Speaker BSides, Roadsec, OWASP
Tsushin Framework (Open-Source)
Origami (Web Security)
Author: Applied Threat Hunting (EC-Council)
Market

AI is transforming Cybersecurity.
Are you ready?

Companies seek professionals who understand both sides — AI and security. This course positions you at the most valued intersection in the market.

Full Coverage

Four Quadrants. Zero Blind Spots.

The most comprehensive AI + Cybersecurity course, covering all 4 dimensions: attack and defense, agentic and programmatic — with hands-on labs in each one.

22+ Exercises 4 Quadrants Red + Blue Agentic + Programmatic
▲ Red Team — Offensive
▲ Blue Team — Defensive
Agentic
▼
Programmatic
Red Agentic Techniques 9
Interactive attacks against/with AI agents
Prompt Injection
SSRF via Tool Invocation
Data Poisoning
Credential Theft via AI
Indirect Prompt Injection
Command Injection via AI
+3 exercises
Blue Agentic Tools 4
Active defense with real-time AI agents
Guardrails (Sentinel, N8N, Custom)
DLP Agent
Prompt Injection Firewall
Behavioral Sandbox
Red Programmatic Tools 4
Automated offensive frameworks powered by AI
Recon & Scanning via MCP
HexStrike: Pentest with 31 Tools
Redamon: 6-Phase Red Team Pipeline
ARTEMIS: Scanner Multi-Agente
Blue Programmatic Techniques 5
Automated defensive analysis pipelines
Semantic CVE Analysis (RAG)
MITRE ATT&CK Mapping
Executive Cross-Report
Threat Intel Multi-Agente (CrewAI)
Automated CVE Lookup

Summary: They teach how to use AI. We teach how to hack AI agents, how to defend them with real guardrails, how to automate pentesting with AI and how to analyze threats with intelligent pipelines. Four dimensions. One course.

Solid Foundations

Understand how LLMs work, advanced prompt engineering, RAG, MCP, and memory architectures — not just “using ChatGPT”.

Security-First

Prompt injection, agent takeover, data poisoning, supply chain attacks — learn to attack and defend AI agents with real labs.

Hands-On Building

Leave the course with chatbots, automation pipelines, multi-step agents, and security reports built by you.

Instructor
Marcos Vinicios Penha

Marcos Vinicios Penha

AI-Powered Cybersecurity and Innovation SME. Professional with an international career and 15+ years combining deep technical expertise and strategic leadership in offensive, defensive, and cyber intelligence capabilities.

Pioneer in Agentic AI for cybersecurity, leveraging frameworks like Agno, CrewAI, and LangChain to orchestrate complex A2A (Agent-to-Agent) workflows. Founder of Tsushin.io — open-source multi-agent orchestration platform with integrated security. Creator of Origami, open-source Swiss Army knife for Web and API security.

Speaker at BSides, Roadsec, and OWASP. Author of the course “Applied Threat Hunting” (EC-Council/CodeRed). MSRC Leaderboard 2022/Q2 (High Severity) for a zero-day in MS Teams.

linkedin.com/in/iamveene
Offensive Security DFIR Detection Engineering Agentic AI Speaker
Course Curriculum

10 Complete Modules

Click any module to see detailed topics. From beginner to advanced, featuring hands-on labs with real-world tools.

01

AI and LLM Fundamentals

21 slides · Concepts, Prompt Engineering, Agents

What is Artificial Intelligence?
How LLMs Work (Transformer)
Key LLM Concepts
State of the Art: 2026
The AI Market in Numbers
SaaS vs Local Models
Prompt Engineering
Zero-Shot / Few-Shot / CoT
Meta Prompting
Assistants vs Agents
Anatomy of an Agent
When to Use Each Approach
02

Architecture and Essential Concepts

20 slides · MCP, RAG, Embeddings, Memória

Model Context Protocol (MCP)
MCP: How It Works
MCP Security
RAG: Complete Pipeline
RAG Applied to Cybersecurity
Databases: 3 Paradigms
Embeddings: Text → Vector
Chunking: Strategies
Knowledge Graphs
Memory Architectures (4 Layers)
Shared Memory: Multi-Agent
03

Frameworks, Tools, and IDEs

22 slides · N8N, CrewAI, Agno, Tsushin

Overview: Decision Matrix
N8N: Visual Automation with AI
CrewAI: Multi-Agent Code-First
Agno: Agents with Tools
Agent Skills for Cyber
AI-Powered IDEs: Cursor Rules & CLAUDE.md
Codex: OpenAI CLI Agent
Antigravity: Agent-First IDE (Google)
Kilo Code: Multi-Mode + CLI
Showcase: Tsushin Framework
Tsushin: Sentinel Security System
04

Hands-On: First Agents

21 slides · N8N, Agno, CrewAI, NotebookLM, Gemini Gems

Environment Setup
N8N: CVE Lookup Agent (Low-Code)
Agno: CVE Threat Intelligence Agent
Automating the Routine
MCP + Cursor: Modern Automation
NotebookLM: Cross-Report Analysis
Meta Prompting with Gemini Gems
CrewAI: CVE Analysis Multi-Agente
Guided Exercise
05

Strategic Decisions

18 slides · When to use AI, Mistakes, HexStrike, Redamon

When to Use / NOT Use LLMs
Decision Tree: LLM or Programmatic?
Hybrid Approach in Production
Common Mistakes in AI + Cyber
Case Study: HexStrike AI (Dual-Use)
Redamon: Red Team Open-Source
AI in Real Attacks (Anthropic 2025)
Implementation Checklist
06

Attacks, Defenses, and Ecosystem Security

34 slides · OWASP, Guardrails, MCP, IDEs, Supply Chain, Shai Hulud 2.0

Attack Surface and OWASP Top 10
Prompt Injection, Agent Takeover, Tool Invocation
Guardrails and Defense in Depth
MCP Security, IDEs, Supply Chain & Shai Hulud 2.0
Principles and Production Checklist
07

Labs: AI Red Team — Agentic

12 labs · Prompt Injection, Credential Theft, SSRF, Data Poisoning, Tool Poisoning, MCP Attacks

Prompt Injection in Agents (N8N vs Tsushin)
Credential Theft via Helpdesk AI
SSRF via Agentic Tool Invocation
Indirect Prompt Injection via CV
Data Poisoning: Cross-Agent via Redis
Rug Pull: Knowledge Base Poisoning Coming Soon
Command Injection via Insecure Eval Coming Soon
Token Passthrough: Confused Deputy Coming Soon
Tool Poisoning: Supply Chain Attack Coming Soon
MCP Server Poisoning: Description Injection Coming Soon
Multi-Chain Agent Injection Coming Soon
Arcanum Taxonomy Attacks (64 payloads) Coming Soon
08

Labs: AI Blue Team — Agentic

7 labs · Guardrails, DLP, Firewall, Sandbox, MCP Gateway, Chain Isolation

Guardrails Sentinel: Agent Defense
DLP Agent: Data Loss Prevention
Prompt Injection Firewall (3 Layers)
Behavioral Sandbox: Policy Engine Coming Soon
MCP Security Gateway: Tool Validation Coming Soon
Agent Chain Isolation: Blast Radius Coming Soon
Taxonomy-Driven Defense (24 regras) Coming Soon
09

Labs: AI Red Team — Programmatic

5 labs · Recon, HexStrike, Redamon, ARTEMIS, PyRIT

Recon and Scanning with AI (Cursor + MCP)
HexStrike AI: 30+ Tools via MCP
Redamon: Agentic Red Team + Neo4j
ARTEMIS: Autonomous Multi-Agent Scanner
PyRIT: Automated Red Teaming (Crescendo, PAIR, TAP) Coming Soon
10

Labs: AI Blue Team — Programmatic

4 labs · RAG/NIST, MITRE ATT&CK, NotebookLM, Benchmarking

Vulnerability Database + RAG Pipeline (330K+ CVEs)
MITRE ATT&CK + Agno: Threat Intelligence
NotebookLM: Cross-Report Synthesis
Automated Defense Benchmarking (ASR Métricas) Coming Soon
Register Now
Market

The Market Is Screaming for This Professional

The numbers don’t lie: mastering AI + Cybersecurity gives you an immediate competitive edge.

$250B+
Global Cybersecurity Market (2025)
Growing 12-14% annually. Companies need professionals who understand AI.
3.5M+
Global Cybersecurity Talent Shortage
The gap widens as AI transforms the attack surface. (ISC2 2024)
3-5x
Estimated Salary Premium
Professionals with hybrid skills (AI + Cyber) command higher salaries.

Sources: ISC2, Gartner, Statista, LinkedIn Economic Graph (2025)

Launch Offer

Invest in Your Career

Special pricing for early students. Limited spots at this rate.

Limited Time Offer
$300
$240
20% OFF — Launch pricing for early students
12h+ video lessons with 10 complete modules (160+ slides)
30+ hands-on labs (~8h practice) with real-world tools
Source code for all course projects
Access to cutting-edge research to transform your career
Certificate of completion
Lifetime access to the full course
Free updates, including new labs
Register Now — Secure Your Spot

Limited Spots at This Rate

The promotional price of $240 (20% OFF) is exclusive to early students. Once spots are filled, the price returns to $300. Register now to secure your place.

Questions

Frequently Asked Questions

Do I need to know how to code?

Basic Python knowledge helps, but it is not required. The course guides you from the fundamentals, and the labs include step-by-step instructions. If you can navigate a terminal, you can follow along.

Is it just theory or is there hands-on practice?

The course has 30+ hands-on labs (~8h practice) with real-world tools: N8N, nuclei, nmap, Agno, CrewAI, Tsushin, PyRIT, and more. You attack and defend AI agents in practice. This is not glorified PowerPoint.

How long do I have access?

Lifetime access. Watch whenever and as many times as you want. Future updates are included at no additional cost.

What if I don't like it?

Unconditional 7-day guarantee. If you don't like it, Hotmart refunds 100% of the amount. No questions asked, no red tape. Zero risk.

How is this course different from other AI courses?

Other courses teach how to use AI as a productivity tool. This one teaches AI security: how to attack agents, how to defend with professional guardrails, how to automate pentesting with AI. No other course does this with hands-on labs and real-world tools.

What payment methods are accepted?

All major credit cards, PayPal, and international methods through our platform. Installments may be available depending on your card issuer.

I already code well. Isn't the course too basic?

No. Modules 6, 7, and 8 cover advanced attacks (prompt injection, agent takeover, data poisoning, SSRF via agentic tools), defenses with professional guardrails, and Red Team and Blue Team labs with tools like HexStrike AI and Redamon. If you already code, you go straight to the good stuff.

Don't wait for the market to change
without you.

The demand for professionals who master AI + Cybersecurity is exploding. Position yourself now with practical, differentiated knowledge.

Secure My Spot for $240
Tsushin Ecosystem

Three Pillars, One Mission

Practical education, open-source technology, and security tools, united by AI protection.

Tsushin.io Academy

Educational Platform

Practical, hands-on training at the intersection of AI + Cybersecurity. Courses, labs, and certificates for professionals who want to master secure AI agents.

Access Course

Tsushin Framework

Open-Source · MIT

Multi-agent framework with Sentinel Security System, 16 skills, semantic memory, and full observability. Free and open-source.

View on GitHub

Origami

Open-Source · Hybrid

Swiss Army knife for Web and API security. Blends programmatic and agentic capabilities for comprehensive offensive testing.

View on GitHub
$300 $240 — Limited Spots
Register Now