Hey Friend,

I hope you are doing well! I am thrilled to announce that I have been nominated for two awards and I need your help to win (It takes less than a min to vote)! 

• 🏆 SANS Difference Makers (Innovation of the Year for NOVA): This is the last day to vote! 
• 🇫🇷🇦🇺 French-Australian Award for Research & Innovation (LCA): You can vote here!

The two nominations recognize my work in AI Security and Threat Intelligence, and honestly, this is a real honor!

If my work has ever helped you or sparked some idea, your vote would mean a lot! I really appreciate it (and feel free to let me know if you do)! ❤️ 

Now, a quick update on what I have recently release.

Adversarial Prompt Taxonomy v0.1
A structured way to classify prompt-based attacks and failures in AI systems. The goal is to help researchers, red teamers, and defenders to speak the same language when investigating or mitigating LLM threats! 

This is a V0 but let me know what you think.

You can access the Adversarial Prompt taxonomy here: https://jupyter.securitybreak.io/IoPC/AdversarialPrompts.html 

Proximity
My new MCP scanner that detects malicious or impersonated Model Context Protocol servers using NOVA rules.

It can spot shady behaviors like hidden BCCs or silent data leaks, exactly what happened with the recent “Postmark MCP” incident.

You can download Proximity here: https://github.com/fr0gger/proximity 

Both are part of the bigger vision behind NOVA, shaping the AI Threat Intelligence foundation for all analysts.

Thank you all for the support, the feedback, and the motivation to keep pushing the field forward. It really means a lot!

Until next time ✌️

Thomas

PS: I just received the confirmation that my training GenAI for CTI will be taught at BlackHat ASIA, stay tuned for the announcement.