Eli Cohen argues that AI-assisted coding increases the need for continuous application-security testing. Static analysis can identify suspicious patterns cheaply, while dynamic checks and human understanding help establish whether a finding matters in the application's actual business context.
Eli Cohen describes an orchestratorAgent orchestration coordinates AI agents, tools, people, tasks, state, and control flow so a larger workflow reaches a verified outcome. that delegates reconnaissance and focused testingAI agent delegation assigns a bounded task, context, authority, and expected result from one participant to an agent or another agent. to specialized agents. A separate judging step checks proposed exploits and gathers reproducible evidence, rather than trusting an agent's first claim or treating every scanner alert as a confirmed issue.
Eli Cohen presents this as a vendor's architecture and discusses evaluation questionsAgent evaluation tests whether an AI agent completes tasks correctly, consistently, and within its required boundaries. around coverage, false positives and useful remediation. The talk is not an independent benchmark, and its promotional comparisons do not establish general superiority.
Watch on YouTube
