The 3 AI Agents everyone should have…

Key Concepts

AI agents, cyber security, phishing, two-factor authentication (2FA), authenticator apps, URL analysis, VirusTotal, URLScan, NA10, virtual private server (VPS), Hostinger, email breach alerts, Have I Been Pwned, password sharing, software updates, VPN, website security auditing, penetration testing (pentesting).

1. Introduction: The Rising Threat of AI-Powered Cyber Attacks

The video addresses the increasing prevalence of AI-powered cyber attacks and their higher success rate compared to traditional methods.
Fact: Over 166 million individuals were affected by cyber attacks this year.
Statistic: AI-powered attacks have a four times higher success rate than traditional phishing.
Example: AI voice spoofing is cited as a new technique where voices are cloned to extract money.
The speaker emphasizes the importance of cyber security, noting that digital lives are becoming increasingly important.
The video introduces an "AI-first cyber security course for 2026" and promises to show free AI agents for cyber security.
The speaker claims that using these AI agents will put viewers ahead of 99% of people in terms of cyber security.

2. Fundamental Cyber Security Practices

The video stresses that most hacks are due to simple mistakes like social engineering and phishing, not sophisticated attacks.
Key Point: Protecting against 95% of cyber attacks doesn't require being a cyber security professional.

2.1. Two-Factor Authentication (2FA) with Authenticator Apps

Problem: SMS-based 2FA is vulnerable to SIM card spoofing.
Solution: Use authenticator apps like Google Authenticator or 1Password Authenticator.
Actionable Advice: Switch from SMS 2FA to a dedicated authenticator app.

2.2. Protection Against Phishing

Definition: Phishing involves fake emails or messages from trusted companies to steal passwords or install malware.
Method: Hover over suspicious links to see the actual redirect URL in the bottom left of the screen.
Example: A recent phishing wave targeted PyPI maintainers with fake "verify your email" messages.
Tool: VirusTotal is a free website for analyzing files and URLs for malicious content.
Process: Before clicking a suspicious link, paste it into VirusTotal.

3. AI Agent for Automated Phishing Analysis

The speaker introduces an AI agent built to automate the analysis of emails for phishing attempts.
Benefit: The AI agent runs continuously, reducing the likelihood of falling for phishing scams.
Platform: The AI agent is built on NA10, a drag-and-drop platform for creating AI agents without programming.

3.1. NA10 Workflow Breakdown

Gmail Trigger: Checks the inbox every minute for new, unchecked emails.
Batch Splitting: Splits email content into batches in case of multiple URLs.
Code Block (Python): Uses regular expressions (regex) to extract URLs from the email body.
- Note: The speaker used Claude to generate the Python script.
If/Else Node: Determines if a URL exists in the email.
HTTP Node: Sends POST requests to VirusTotal and URLScan APIs to check URLs for malicious content.
Error Protection & Delay: Includes error handling and a 60-second delay for URLScan due to its occasional flakiness.
Combine Node: Merges reports from VirusTotal and URLScan.
Data Cleaning (JavaScript): Cleans the data for the AI agent.
- Note: The speaker used ChatGPT or Claude to generate the JavaScript code.
AI Agent (GPT-4 Mini): The "brain" of the automation, analyzing the data and making decisions.
- System Prompt: Analyzes results from VirusTotal and URLScan, raw URL, and raw email data to determine if the email is safe, suspicious, or malicious.
Gmail Send: Sends a notification email based on the AI agent's decision.

3.2. AI Agent Customization

The AI agent can be configured to send notifications via Telegram, Discord, or WhatsApp.
The speaker suggests using Anthropic's Claude 4.5 model for a more powerful AI agent (and encourages viewers to request a video on this).

4. Hosting AI Agents on a Virtual Private Server (VPS)

Importance: AI agents are only useful if they run 24/7.
Solution: Host AI agents on a VPS.
Recommendation: Hostinger is recommended for its optimized VPS for NA10 and simple setup.

4.1. Hostinger Setup Process

Visit the Hostinger site (link in video description).
Choose a VPS plan (KVN2 recommended).
Select a billing period (24 months recommended for long-term use and better deals).
Use coupon code "David" for an additional 10% off.
Select the server location closest to you.
Choose the operating system (NA10).
Complete the checkout process.
Access and manage the VPS through the Hostinger panel.
Manage AI agents through the NA10 server.

5. Additional Cyber Security Lessons

5.1. Using Multiple Emails

Problem: Using a single email for everything increases vulnerability.
Solution: Separate sensitive accounts (banking, finance) from public communication emails.
Example: Use one email for job applications and another for banking apps.

5.2. Updating Software

Importance: Software updates include security fixes.
Actionable Advice: Enable auto-updates on sensitive apps.

5.3. Email Breach Alerts

Problem: Email breaches can compromise accounts and lead to identity theft.
Solution: Build an AI agent that monitors emails for inclusion in recent breaches using "Have I Been Pwned."
Example: A Canadian university lost $10 million and a Toyota supplier lost $37 million through email scams.

5.4. Password Sharing

Problem: Sharing passwords increases the attack surface.
Actionable Advice: Minimize password sharing.

5.5. Uninstalling Apps

Importance: Reducing the number of installed apps reduces the attack surface.
Actionable Advice: Uninstall unused apps (3+ months).

5.6. Auditing Websites (Especially AI-Coded Sites)

Problem: AI-coded websites may have security vulnerabilities.
Solution: Use an AI agent to regularly pentest the website.
Example: The T app suffered a massive breach due to an exposed Firebase bucket.

5.7. Using a VPN

Benefit: Hides your IP address and adds a layer of security, especially on public Wi-Fi.
Pro Tip: Research the ownership of VPN companies.

5.8. Physical Security

Taping Ports: Prevents unauthorized USB devices from being inserted.
Camera Cover: Provides physical protection for the webcam.

6. AI Agent for Email Breach Monitoring

Schedule Trigger: Runs daily at 9:00 AM.
Google Sheets: Retrieves a list of emails from a Google Sheet.
Have I Been Pwned API: Checks each email against the Have I Been Pwned database.
AI Agent: Analyzes the report and notifies the user if an email has been compromised.

7. AI Agent for Website Security Auditing

Trigger: Landing page URL.
Web Scraping: Scrapes the content of the website.
Header Extraction: Extracts the headers of the website.
Security AI Agent (Headers): Analyzes the headers for leaking metadata or other issues.
Security AI Agent (Website Structure): Analyzes the website structure for potential vulnerabilities.
Merge Node: Merges the outputs of the two agents.
Data Processing & HTML Conversion: Processes the data and converts it to HTML for email notification.
Email Notification: Sends an email with the security audit results.

8. Conclusion

The video provides a comprehensive overview of cyber security fundamentals and demonstrates how AI agents can be used to automate and enhance security measures. It emphasizes the importance of proactive security practices, such as using authenticator apps, being cautious of phishing attempts, and keeping software updated. The video also highlights the benefits of using AI agents to automate tasks like phishing analysis and email breach monitoring, making cyber security more accessible to non-technical users. The speaker encourages viewers to implement these strategies and utilize the provided AI agent workflows to improve their overall cyber security posture.