We find the gaps in your defenses before someone with bad intentions does. Real attack techniques, honest reporting, and no fluff.
We run the same attacks real adversaries use, against your actual environment. No checkbox audits, no generic reports.
Starting from zero access, we go after your most valuable assets using the same tools and tactics real attackers use. We're testing whether your people, your processes, and your technology can actually catch us in time.
Most cloud breaches happen because of misconfigurations, over-permissioned accounts, and forgotten resources — not zero-days. We look at your cloud the way an attacker would, and connect the dots between low-risk items that automated scanners treat as separate.
MFA is not the end of the conversation. We run phishing campaigns using the same setups threat actors actually use, including proxies that can bypass MFA entirely. After every campaign you get real numbers: who clicked, who submitted credentials, and what to do about it.
Automated scanners produce a lot of noise and miss the things that actually matter. We pair tooling with manual review by people who understand how code gets written and how it gets broken, which is how we find the logic flaws and authorization bypasses that tools walk straight past.
Modern applications have a lot of moving parts: multiple frontends, microservices, third-party integrations, public APIs. We test the whole picture manually, connecting findings that look minor in isolation into realistic scenarios that show you the actual business impact.
From your public perimeter to your internal domain controllers, we map and exploit your infrastructure the way a real attacker would. Every exposed asset gets looked at, exploitable vulnerabilities get prioritized, and you see the full impact of what a network-level breach actually looks like.
This is not a checkbox security firm. Every engagement is led by a practitioner who has spent years developing offensive tooling, getting past enterprise EDRs, and exploiting real production environments.
We have presented at international security conferences, contributed research to the offensive security community, and hold the certifications that actually require you to break things under exam conditions.
Red Team Security SRL was founded by Andrei Grigoras — a self-driven, self-taught security professional with over 7 years of hands-on offensive security experience. Starting from a Junior Security Researcher role in 2018, Andrei progressed rapidly through SOC analysis, consulting at Accenture, senior penetration testing at TwelveSec and Casumo, independent tooling research, and most recently as a Red Team Operator at PwC Ireland — leading engagements against organizations with 5,000+ employees.
Andrei is a recognized speaker at international security conferences including DefCamp, BSides Athens, BSides Tirana, HEK.SI, and INFOSEK.SI. He has developed multiple custom offensive tools used in real engagements, contributed original research to the community, and competed for Team Romania in the European Cyber Security Challenge.
Leading red team engagements for large enterprises (5,000+ employees), developing custom attack infrastructure, modifying tools like Kerbrute and SOAPHound, and supporting External Attack Surface Management.
Built Web Application Vulnerability Scanner MVP with scraper, fuzzer, and AI-powered anomaly detector. Developed Automated Red Team engine with 60+ CVE PoCs and Terraform-based infrastructure.
Led red team operations using Mythic and Cobalt Strike, deployed resilient red team infrastructure, mentored junior staff, and conducted mobile, web, infrastructure, and phishing assessments.
Security testing across iGaming products including code review, external/internal PT, phishing, and infrastructure assessments. Close cooperation with dev/ops teams.
Security training development, SAST/DAST assessments at Accenture. SIEM log investigation and MDR endpoint analysis at SecureWorks.
Researched new vulnerabilities, implemented them in the platform, and participated in bug bounty programs.
Industry-recognized offensive security certifications backed by real-world competition wins and community contributions.
Compromised 80+ machines on HackTheBox including Rastalabas challenges, maintaining Pro Hacker rank through consistent offensive security practice.
Here is how a typical engagement runs, from first contact to the final fix verification.
We agree on clear boundaries, objectives, and what success looks like. Scope, timeline, communication channels, and emergency contacts are locked in before anything starts.
We map your external attack surface using the same sources an attacker would: domains, IP ranges, employee data, technology stack, supplier relationships, and any credentials that have leaked publicly.
We exploit what we find and go as far as the scope allows. Findings get chained together to show what the real damage would be, not just a list of theoretical risks.
You get a technical report for your security team with full reproduction steps, and an executive summary your management can actually read. We also do a live debrief to walk through findings together.
We are available while your team works through fixes, and we re-test the patched findings at no extra charge to confirm everything is properly resolved.
We don't rely solely on public tools. Our operator develops and maintains a suite of custom offensive capabilities.
A heavily modified evilginx3 fork featuring dynamic HTML obfuscation to defeat browser-side static scanning engines, link cloaking, and real-time credential capture piped into Gophish campaign metrics.
A feature-extended Gophish fork with native QR code campaign support, full bidirectional integration with the custom Evilginx3 for AiTM flows, and improved credential export and reporting functionality.
Custom multi-stage DAST tool: a Selenium-driven scraper maps all inputs, a fuzzer tests XSS, SSTI, LFI, SQLi, and open-redirect vectors, and an AI-assisted anomaly detector reduces false positives automatically.
Automated tool that submits phishing domains for recategorisation across multiple vendor reputation platforms using Selenium and an integrated captcha-solving engine — keeping infrastructure off blocklists.
60+ hand-written RCE exploit proof-of-concepts covering CVEs from the last 5 years. Each includes a reproducible vulnerable environment, cross-compiled Linux/Windows binaries, and documented exploitation steps.
Two-variant ransomware PoC developed for red team engagements: a full-UI simulator that replicates real-world ransomware UX and workflow for executive-level impact demos, and a fileless in-memory variant that operates without touching disk to evade EDR-protected environments.
Every engagement starts with a confidential scoping call where we talk through your environment, objectives, and what you want to get out of the test. Reach out and we will get back to you within 24 hours.