Sentry Cybersecurity

Sentry Cybersecurity Sentry provides cutting edge cyber security service

Founded in 2016, the company is established in SEE - with a presence in the United States comprised of Specialized Engineers, Former Military / Navy SEALS Personnel, and Former US Government Officials.

Tool and function calling are now the action layer of modern systems. That is where the interesting exploitation happe...
02/06/2026

Tool and function calling are now the action layer of modern systems.
That is where the interesting exploitation happens, and it is where the security work needs to be.

29/05/2026

More context? Read "Cracking Password Reset Mechanisms"

LinkInBio.

More than 2,500 pe*******on tests over the last decade taught us one thing: what gets you breached is rarely the vulnera...
26/05/2026

More than 2,500 pe*******on tests over the last decade taught us one thing: what gets you breached is rarely the vulnerability your scanner flagged.
It's the forgotten subdomain, the undocumented API, the cloud role that looked fine on its own.

That's what our team does every day. Human-led testing across your whole stack, with findings you can actually fix.

♘♞ See more → https://lnkd.in/dC-qhMUd

22/05/2026

More context? Read "LLM API Misconfigurations"

LinkInBio.

No technical blog today. Just a hacker music playlist from
19/05/2026

No technical blog today. Just a hacker music playlist from

For YOU who said "I'll read it later" and then didn't:"Semantic vs. Token-Based LLM Injections", repurposed, shorter, an...
14/05/2026

For YOU who said "I'll read it later" and then didn't:
"Semantic vs. Token-Based LLM Injections", repurposed, shorter, and designed.

Once again:
♞ Semantic injections manipulate meaning.
♘ Token-based injections exploit the layer beneath language: reserved delimiters, gradient-optimized suffixes, tokenizer mismatches between safety classifiers and target models.

Swipe through. LinkInBio

If you’re building with , you’re going to thank Armend Gashi for this VERY detailed blog piece, worth reading before at...
06/05/2026

If you’re building with , you’re going to thank Armend Gashi for this VERY detailed blog piece, worth reading before attackers do.

What happens when untrusted users can touch parts of your LLM API that should belong only to the server?
System roles. Tool calls. Model parameters. Token limits. Reasoning controls.
LinkInBIO

This article describes the process that uncovered a prefix oracle in . The goal was to provide a complete path for iden...
16/03/2026

This article describes the process that uncovered a prefix oracle in .
The goal was to provide a complete path for identifying, validating, and exploiting side-channels across any web system.

♞ Continue reading: Link in the comments.

27/02/2026

And for the last cropped piece from the podcast with Walter Haydock, we’ll leave the audience with our two favorite educational resources:

♞ Your weekly dose of everything you need to know about AI developments, emerging trends, and tactics in , coming from for our audience, and more.

♞ Risk Readiness Kit. If you're building RAG apps, agents, copilots, or tool-calling systems, this will save you time.

Links in comments.

The trilemma describes a tension between three properties we increasingly want from systems:♞ Generality♞ Agency♞ Alig...
25/02/2026

The trilemma describes a tension between three properties we increasingly want from systems:
♞ Generality
♞ Agency
♞ Alignment

Read the full blog (link in the comment) and let us know what's ur take on this.

Address

Pristina
10000

Opening Hours

Monday 08:00 - 20:00
Tuesday 08:00 - 20:00
Wednesday 08:00 - 20:00
Thursday 08:00 - 20:00
Friday 08:00 - 20:00

Telephone

+38349686668

Website

Alerts

Be the first to know and let us send you an email when Sentry Cybersecurity posts news and promotions. Your email address will not be used for any other purpose, and you can unsubscribe at any time.

Contact The Business

Send a message to Sentry Cybersecurity:

Shortcuts

Share

Category