The Future Of Augmented Appsec Integrating Llm Reasoning Into Burp Workflows With Vista

This talk addresses the challenges faced by penetration testers, such as passive blindness, payload crafting overhead, and expertise fragmentation, and explores how artificial intelligence can be integrated into their workflows to overcome these issues. Led by Rishav Raj and Rajkumar Rathod from Fidelity Information Services, the session features VISTA, an open-source Burp Suite extension that utilizes context-aware LLM reasoning to analyze traffic and surface findings without undermining human judgment. The discussion highlights VISTA's capabilities in identifying hard-coded secrets, hidden parameters, and exploitable endpoints, as well as the use of customizable system prompts and a payload library that tailor to specific methodologies and workflows. Additionally, the session examines how VISTA mitigates limitations of LLMs, such as context loss and hallucination, through effective session memory management, scoped analysis, and model flexibility across various AI platforms.