Lufsec

Information Security Consulting and Training. Check out our online security courses! 💻

[Lufsec @Lufsec]

“Just read what they say.”

Dmitri Alperovitch explains why the idea that trade would change China was an illusion.

Take IP. Build competitors. Push Western companies out. Compete globally.

EVs. Batteries. Solar. Telecom. Rare earths. AI.

Full episode on LufSec YouTube.

[Lufsec @Lufsec]

Operation Aurora changed cybersecurity.

Dmitri Alperovitch explains how the Google hack helped the industry realize nation-states were not just targeting governments — they were targeting private companies too.

Full episode: https://t.co/JdrX39mFJs

#Cybersecurity

[Lufsec @Lufsec]

“We don’t actually have a cyber problem. We have a China, Russia, Iran, North Korea problem.”

Dmitri Alperovitch explains why cybersecurity is now inseparable from geopolitics, nation-states, criminal safe havens, and ransomware.

Full episode: https://t.co/JdrX39mFJs

[Lufsec @Lufsec]

Inside Cyber Minds is back.

Season 2 starts June 10 at 4 PM with Dmitri Alperovitch as the first guest.

This season: Sabu, Katie Moussouris, Roger Grimes, Toni de la Fuente, JP Calabio, Marcio Cots, David Garvin, and more.

Subscribe to LufSec.

#Cybersecurity

[Lufsec @Lufsec]

Secure boot won’t save a blind IoT fleet. Attackers usually get in through exposed management, weak OTA logic, bad remote access, and forgotten edge devices long before firmware integrity matters. New breakdown:https://t.co/C81fsrqBXm

[Lufsec @Lufsec]

New documentary live: The Dark Web Uncovered.

How do stolen data markets, phishing kits, ransomware crews, and crypto laundering really work? This breaks down the dark web’s role in modern cybercrime.

Watch: https://t.co/yAKJV94AUH

#Cybercrime #DarkWeb #Cybersecurity

[Lufsec @Lufsec]

If your first security boundary is the vehicle gateway, you may already be late.

Modern car hacking now starts in apps, APIs, telematics, and account logic before the CAN bus ever matters.

More: https://t.co/p2lmKoPMmp
Course: https://t.co/ZmLQZo2MpB

[Lufsec @Lufsec]

One poisoned MCP server can own your AI agent.

If your agent trusts the wrong tool, prompt injection becomes a breach path.

https://t.co/vnlO4BEzn5
https://t.co/BHBdgG7dDO

[Lufsec @Lufsec]

One weak vehicle API can make car hacking remote.

If apps, telematics, or backend APIs trust too much, attackers may never need to touch the bus.

Tomorrow on LufSec:
https://t.co/JunsL9Wt6t

#CarHacking #AutomotiveCybersecurity

[Lufsec @Lufsec]

One forgotten edge device can sink your IoT security.

CISA’s edge-device push is a good reminder: attackers do not need your cleanest firmware bug if an old router, gateway, or camera is still exposed.

New post:
https://t.co/XUOj03ryIG

#IoTSecurity #IoTHacking

[Lufsec @Lufsec]

One PR comment can make an AI agent leak secrets.

If untrusted content can reach tools, tokens, or workflows, prompt injection is no longer a prompting bug. It’s a security architecture problem.

New post:
https://t.co/rhDgJ9dolj

#AIHacking #PromptInjection

[Lufsec @Lufsec]

Most vehicle security still trusts messages that were never designed to prove identity.

That’s why CAN bus defenses still fail: the real issue is architectural trust, not just monitoring.

New post:
https://t.co/UxaIUkHwWs

[Lufsec @Lufsec]

Prompt injection becomes a breach path when AI agents can read untrusted content and use tools. Trust boundaries, least privilege, validation, and approval gates matter.
https://t.co/BHBdgG7dDO
#AIHacking

[Lufsec @Lufsec]

A vehicle gateway is not a security boundary just because it sits between networks.

If one compromised ECU can still pivot through permissive forwarding, weak diagnostics, or poor traffic policy, the boundary is mostly theater.

Read: https://t.co/hAC4YS78Pn

#CarHacking #lufsec

[Lufsec @Lufsec]

Prompt injection isn’t just an LLM bug. It’s a trust-boundary failure.

If untrusted emails, PDFs, webpages, tickets, or RAG context can steer an agent, the real issue is what your system trusted too cheaply.

https://t.co/BHBdgG7dDO

#PromptInjection #AISecurity #AIAgents

[Lufsec @Lufsec]

Smart devices are endpoints.

Cameras, routers, locks, wearables, and sensors can become attack paths through default passwords, unpatched firmware, open ports, weak Wi-Fi, or insecure APIs.

If it connects to the internet, secure it.

#IoTSecurity #CyberSecurity #LufSec

[Lufsec @Lufsec]

AI will move fast.
Governance decides whether it moves safely.
This is where real AI risk lives.

#AIGovernance #AISecurity #ResponsibleAI #CyberTok #EnterpriseAI

[Lufsec @Lufsec]

🚨 The ISC2 2025 Cybersecurity Hiring Report just exposed the truth:
➡️ 90% of hiring managers value certifications & hands-on skills over degrees.
➡️ Soft skills now outrank tech skills.
The future of cybersecurity belongs to learners who act — not just study.

[Lufsec @Lufsec]

⚔️ The AI battlefield is here.
On Oct 9, I’ll unpack:
• Prompt injection & data poisoning attacks
• Real AI exploits
• Strategies to secure AI deployments
📅 Oct 9 | 10:00 AM EDT
🔗 Register:https://t.co/4KndDnezpD

#AISecurity #Cybersecurity #GenAI #LLMSecurity

[Lufsec @Lufsec]

🎉 My First Cybersecurity Conference in Florida: BSides Orlando 2025
https://t.co/JIlFzmuZVR