General
Security Researcher & AI Red Teamer
Onurcan Genç
I find vulnerabilities in web applications and language models, then publish the findings so defenders can act on them. 8 CVEs disclosed, 14+ months of hands-on pentesting, OWASP GenAI contributor.
BlueDot Impact — Technical AI SafetyOWASP GenAI Red Team Lab ContributoreWPTXv3eWPTv2C-AI/MLPenC-AgAIPenCompTIA Security+
CVE Disclosures
CVE-2026-34156
CVSS 10.0NocoBase VM Sandbox Escape to RCE
NocoBase v2.0.26
VM sandbox escape through console prototype chain in Workflow Script Node. Authenticated attacker achieves RCE as root inside Docker containers — database credential theft, arbitrary file access, reverse shell.
CVE-2026-9558
CVSS 9.9Mautic SSTI via Twig Themes to RCE
Mautic 5.2.10
Server-side template injection through theme upload. Low-privilege Designer user with core:themes:create achieves remote code execution via unsandboxed Twig rendering.
CVE-2026-3251
HighStored XSS
Stored cross-site scripting vulnerability. Reported through T.C. Siber Güvenlik Başkanlığı.
CVE-2025-60506
HighStored XSS in Moodle PDF Annotator
Moodle PDF Annotator v1.5r9
Student-level attacker injects JavaScript via Public Comments. Payload fires when any user views the annotated PDF — session hijacking, credential theft.
CVE-2025-60511
MediumIDOR in Moodle OpenAI Chat Block
Moodle OpenAI Chat Block 3.0.1
Authenticated student impersonates another user (admin, teacher) and sends OpenAI queries using their configuration. API misuse and admin-data exposure.
CVE-2025-60507
HighReflected & Stored XSS via PDF Upload
Moodle GeniAI Plugin 2.3.6
Teacher uploads PDF with embedded JavaScript. The assistant outputs an unsanitized HTML link — payload executes for any user who clicks it.
CVE-2025-10228
HighSession Fixation
Session fixation vulnerability. Reported through USOM.
CVE-2025-57520
HighStored XSS in Decap CMS
Decap CMS ≤ 3.8.3
Contributor/editor injects payloads into title, tags, description, and body fields. Fires when admin opens content preview — session hijacking, arbitrary JS execution.
Experience
Jul 2026 – Present
HUX AI
London, UK · Remote
AI Research Intern — LLM Guardrails Evaluation
- LLM guardrail evaluation and adversarial testing
- Prompt injection and jailbreak resistance assessment
- AI safety evaluation workflow development
Nov 2025 – Feb 2026
Deloitte
Ankara · Remote
Cybersecurity Intern
- Pentest report QA aligned with MITRE ATT&CK and CVSS
- Severity scoring validation across deliverables
- CTF competition — 1st place
Jan 2025 – Sep 2025
Bilishim Cyber Security
Ankara · Hybrid
Penetration Tester
- Web, mobile, wireless, and LAN penetration testing
- AI/ML red team engagements
- Threat intelligence: dark web leak analysis
- Mentored junior interns
Open Source & Research Contributions
OWASP GenAI Security Project
Multi-Technique Guardrail Bypass Evaluation
Five-stage evaluation sequence across four bypass technique families against a production chatbot. Demonstrates that patching individual bypasses fails — the same intent gets through via different surface transformations.
OWASP GenAI Security Project
Multi-Turn Safety Bypass & System Role Override
Six-stage attack chain: control token injection, role-label spoofing, role-switching disclosure, Crescendo escalation. Mapped to LLM01, LLM02, LLM05, LLM07. Responsibly disclosed via MITRE.
Independent Research
Mechanistic Interpretability — Refusal in Open-Weight Models
Characterizing how refusal is represented in the residual stream of Qwen 2.5. Using targeted interventions to identify which directions are causally responsible for refusal vs. compliance.
Certifications
eWPTXv3
INE
eWPTv2
INE
C-AI/MLPen
SecOps Group
C-AgAIPen
SecOps Group
CompTIA Security+
CompTIA
CNSP
SecOps Group
AWS Cloud Practitioner
AWS
Technical AI Safety
BlueDot Impact
Future of AI
BlueDot Impact
Education
Bilkent University
B.ASc, Information Systems & Technologies
2020 – 2026 · 50% Scholarship
Best Intern AwardSocial Excellence Award
C, C++, Python, Java, PHP, Kotlin, C#/.NET, Oracle SQL/PL-SQL. GoF design patterns, RESTful microservices, algorithmic analysis, IS auditing.
Training Programs
- Boğaziçi Cybersecurity Program (BUSIBER)
- Türk Telekom Cybersecurity Camp
- Deloitte CyberOps Bootcamp — CTF 1st Place
Research Areas
Vulnerability Research
Web framework audits, CVE disclosure, root-cause analysis.
Explore →AI Security
LLM guardrail evaluation, jailbreak methods, mechanistic interpretability.
Explore →CTF Writeups
HackTheBox, TryHackMe, and competition walkthroughs.
Explore →General
Certifications, career notes, and industry observations.
Explore →