Talk All Audiences 13:30 - 14:00 August 07, 2026

Bar Kaduri

Lidan

AI agents have quickly become a core part of production infrastructure. Autonomous systems that invoke tools, execute code, orchestrate workflows, and interact with external services are now deployed across developer tooling, automation platforms, data pipelines, and enterprise environments at scale.

Yet despite the speed of adoption, almost nothing is known about how these systems are actually built or secured in the wild.

To answer that question, we analyzed more than 86K public repos implementing agent logic across LangChain, LangGraph, CrewAI, AutoGen, and MCP, examining prompt construction, tool implementations, authentication models, and permission boundaries. We paired this with internet-wide infra measurement using Shodan, Censys & ShadowServer, surfacing more than 700K exposed agent-related systems, including Ollama inference servers, Ray clusters, n8n platforms & MCP tool servers, many with little or no authentication and numerous known high-impact CVEs.

Bar Kaduri

Head of Research, Capsule Security

Bar Kaduri is a cybersecurity researcher, leader, and international speaker with over 14 years of experience in cloud security, software supply-chain risk, and emerging AI threats. With hands-on expertise in evaluating and stress-testing AI systems, Bar focuses on building practical, resilient defenses that anticipate risks before they are widely recognized. Bar regularly speaks at industry conferences, delivering clear, research-driven insights that bridge deep technical expertise with strategic security leadership in an AI-driven world


Lidan

CTO and Co-Founder, Capsule Security

Lidan has been programming since childhood, driven by a deep passion for data and AI. He previously served as VP of R&D at SecuredTouch, where he helped pioneer behavioral biometrics. Following the company’s acquisition by Ping Identity, the technology he led became a core component of Ping’s SaaS offering. He later joined Transmit Security, where he built their Identity Threat Detection and Response platform from the ground up, targeting large-tier financial organizations. He scaled the team to more than 30 engineers and data scientists, drove ARR from $0 to $20M, and delivered multiple presentations at leading cybersecurity conferences. In 2025, he co-founded Capsule Security, where he serves as CTO. Capsule addresses one of the most critical challenges in modern cybersecurity: securing AI agents. The company provides end-to-end visibility, risk analysis, threat detection, and runtime protection with advanced detection and response.