Prompt Leakage and System Prompt Extraction in Commercial LLM Applications
Two-turn attacks raise prompt leakage success rates to 86 percent across leading models.
Svetlana Prochazka
Staff Researcher & Writer
Svetlana Prochazka
A former cryptographic engineer who worked on protocol design for a European fintech firm, Svetlana now translates dense cryptographic theory into actionable analysis for security practitioners and developers alike.
6 stories
Timing Side-Channel Vulnerabilities in Authentication Code
Attackers can recover secrets one character at a time by measuring response delays.
Svetlana Prochazka
Password Reset Flow Vulnerabilities in SaaS Applications
Attackers exploit weak reset flows to bypass MFA and take over accounts without cracking passwords.
Svetlana Prochazka
Multi-Factor Authentication Bypass Techniques
Attackers now bypass MFA by stealing session tokens after users legitimately log in.
Svetlana Prochazka
JWT Algorithm Confusion Attacks in Production APIs
Attackers forge tokens by controlling which verification algorithm runs.
Svetlana Prochazka
Session Fixation and Session Hijacking in Modern Web Apps
Stolen session tokens now beat passwords as the primary credential attack.
Svetlana Prochazka