Roberto Marc🏹

Researcher @syhunt, monitoring the dark web and researching #opensource software

[DaragonTech @DaragonTechHQ]

Proud to have partnered with Syhunt in developing its next-generation on-device AI technologies, including proprietary AI models, a breakthrough Symbiotic AI security engine, and AI-powered vulnerability analysis

https://t.co/D04zxDJAB3

[Syhunt @syhunt]

🚀 Syhunt is bringing next-generation, on-device AI to application security!

Join us at Security Leaders Brazil 2026 for an exclusive preview of NX, our new AI-native and AI-augmented security product line.

📍 São Paulo | October 21–22
🔗 Learn more: https://t.co/e7QLe45tCp

[Syhunt @syhunt]

DaragonTech started out as Syhunt's AI research division and is now an independent company.

Proud to see what it's shipping.

https://t.co/6AxJ4xz3ni #OpenSource

[Victoria Brandt @veebrandt]

welcome LibLayaX ✨🚀

[DaragonTech @DaragonTechHQ]

AI decisions without an API bill

LibLayaX is an opensource alternative to the Jev API: a shared library that runs the Laya typed-decision model inside your application. Your data never leaves the machine

Win11, Linux, macOS. CPU or GPU

https://t.co/0JX3aFcH6B

#OpenSource #Jev

[DaragonTech @DaragonTechHQ]

LibLayaX 1.0.15 is out: an open-source alternative to the Jev API. It runs Laya decision models locally, from any language, on CPU or GPU.

New:
- a model in one .tar file
- GPU under Wine, Apple Silicon included

https://t.co/0JX3aFcH6B

#OpenSource #AI #ML #LocalAI #Jev

[DaragonTech @DaragonTechHQ]

We just ported the Lua 5.1 library to pure Rust 🦀

Rulua, binary-compatible with lua5.1.dll / .so / .dylib.

Existing apps and C modules just load it.

Try it, break it, tell us 👇
https://t.co/5cVXkbbHsa

#lua #rust #rustlang #coding #opensource

[DaragonTech @DaragonTechHQ]

🚀 Lude 1.0 is here — Lua 5.1.4, ported to pure Delphi 10.2+ & Free Pascal.

A drop-in replacement for lua5.1.dll/.so/.dylib, with the same 123 exports.

🌍 Win, Linux & Mac
💻 x86 & ARM
✅ Passes the official Lua test suite

https://t.co/w9OvBtJPiR

#lualang #pascal #delphi

[Felipe Daragon @FelipeDaragon]

I created MIT-Human license: basically MIT, but AI isn't allowed to help develop the code.

I tested it with Claude. Found a loophole. I patched the license. Now it refuses.

Your turn - give it to your fav coding AI: https://t.co/PmWf0cDqQ2
More: https://t.co/ZUW2TQA4mN

#AI

[DaragonTech @DaragonTechHQ]

One QR code was never enough. So we stacked three.

TriLayerQR: a photo, a name and a hidden memo in one scannable square. A new kind of QR for the AI age.

Free and open source. 👇
https://t.co/YRTBS3ZbBn

#AI #genai #opensource #innovations

[Syhunt @syhunt]

In the Age of AI, Vision Is the Last Competitive Advantage

In his latest article, @felipedaragon explores what that means for leadership in the AI era — and why knowing what NOT to build may be the most important skill of the #AI era.

Read the article: https://t.co/iCguKEzTud

[Syhunt @syhunt]

The Predictions That Came How AI Has Already Transformed Cybersecurity

Read the article: https://t.co/d4I5FzUmtC

#cybersecurity #predictions #prediction #trends #forecasts #infosec #AI #artificialintelligence #agentic

[Syhunt @syhunt]

The Rise of AI Pentesters — and Why Specialization Will Define the Winners

Read the article: https://t.co/tgqAVAjvQM

#aipentester #ai #cybersecurity #agentic #pentesting #infosec #artificialintelligence #genai

[Syhunt @syhunt]

Introducing Cross-Model Scripting (XMS) vulnerabilities:

Our new paper explores cross-site scripting (XSS) risks arising from unsanitized LLM-generated output and introduces the term Cross-Model Scripting (XMS) to describe these AI-mediated injection scenarios

📄: https://t.co/5TxbZw67uG

#paper #cybersecurity #infosec #xss #pentesting #llms #ai #genai

[DaragonTech @DaragonTechHQ]

Introducing CrossSpeak!

A new writing layer designed to bridge the gap between human creativity and AI comprehension. 🤯

It looks different, reads the same. AI still gets it. Pure semantic magic!

https://t.co/K5f9kVaFRk

#llms #ai #genai

[Felipe Daragon @FelipeDaragon]

NEW PAPER: Cross-Model Scripting (XMS):

Check out our new paper that explores AI-mediated injection and our CrossSpeak tool

https://t.co/LgjADW0L2K

[Roberto Marc🏹 @RobertoMarc15]

RT @syhunt: Introducing Cross-Model Scripting (XMS) vulnerabilities - where LLM-generated content gets weaponized 💥.

This new breed of th…

[Syhunt @syhunt]

Tired of chasing vulnerabilities?
Syhunt Code hunts them down. 🤘

Our on-premise static analysis pinpoints vulnerabilities like XSS, SQL injection & more in your web app source code - no manual sifting needed.

Get back to building faster. 😎

#sast #infosec #appsec

[Roberto Marc🏹 @RobertoMarc15]

RT @syhunt: Looks Japanese 🇯🇵, means something else 🤖, unlocks the truth 🔓

Our new project KanaShift derives a keystream from the password…

[Syhunt @syhunt]

Syhunt is thrilled to announce that our @MechasAI division has spun off into an independent European company!

We're excited to continue our partnership as strategic allies, driving innovation in the AI cybersecurity space.

Dive Deeper: https://t.co/rJ5CnxAwkx

#AI #Cybersecurity #Innovation