Delivered a keynote presentation at SAI RESET IT 2025 in Brussels 🇧🇪 on the future of Application Security and Agentic AI. The talk looks at the past, the present and the future of how we build applications and how we integrate security, with a little twist at the end.
Ready to hack and fix LLM applications? This workshop will take you through the OWASP LLM Top 10 via hands-on hacking labs. You will learn various techniques to exploit LLM applications vulnerabilities, and how to implement robust secure design patterns. Whether you are a builder or breaker of LLM applications this is for you! At the end of the day there will be an exciting attack-and- defense wargame. Plenty of learning and fun, guaranteed.
As open source large language models (LLMs) continue to gain traction, so too does the risk of supply chain attacks targeting platforms like Hugging Face. This talk explores the current threat landscape surrounding open source LLMs, focusing on how malicious actors can inject backdoors into models and associated code. We demonstrate practical techniques for bypassing Hugging Face’s existing security controls, revealing how seemingly benign models can conceal harmful behaviors.
GitArmor is an open source tool that intuitively transforms the security requirements and controls for your DevOps implementation into policies as code and enables you to run the checks against your GitHub environment.
Here’s how GitArmor can be a game changer for you:
Policy as Code - Transform your DevOps platform security policies into GitArmor yml files. Stored centrally, these policies can be enforced with on-demand or periodic checks across your GitHub Organization, helping you pinpoint and prioritize areas for improvement.
Security Assessments - Perfect for Security Teams, GitArmor facilitates the reconnaissance phase of possible misconfigurations of the SCM environment.
Dev Team Setup - Ideal for startups, a small development team can utilize GitArmor along with the default policy to ensure their GitHub repositories and organization are securely configured.
After spending the last 1 or 2 years getting your DevOps process right, here it comes the new security guy: "We need to move to DevSecOps". This talk wants to share my personal experience, challenges, and successes as DevSecOps Architect in implementing DevSecOps in different DevOps processes. The talk starts with the main question: "Where do we start?" to then moves to topics like IaC security, the policy as code, SAST, SCA, SBOM, Security Champions, CI/CD security, supply chain security, logging and monitoring, and DevSecOps maturity. Don't look at it as a list but as a mix of connected resources that will increase automation and reduce manual bottlenecks. At the end of the talk, attendees should already be able to picture their DevSecOps journey ahead.
"Attacking and Defending GraphQL Applications: A Hands-On Approach" is a workshop that provides practical knowledge on securing GraphQL applications. Participants engage in hands-on activities to identify, exploit, and mitigate vulnerabilities in GraphQL APIs. The workshop covers attack vectors specific to GraphQL and teaches defensive techniques such as authentication, input validation, and rate limiting.
We presented a new tool to automate cybersquatting attacks and indentify available websites to use in phishing campaigns. The tool also allows you to monitor existing domains for expiration date and buy it as soon as it becomes available. And .... it is opensource
Davide Cioccia & Stefan Petrushevski @ BlackHat Arsenal London 2018
We presented a CLI tool to exploit, detect and prevent cybersquatting attacks. The tool is designed to be fast and give a quick feedback to companies and pentesters.
This talk introduces a new process and practical solution that achieves this – automation of mobile security tests. We are using a combination of existing penetration testing frameworks (Drozer and Needle), UI automation, underlying system commands available in the mobile OS for execution of tests and describe (write) tests in BDD fashion. In this way, you can cover all kind of security tests, such as testing for not encrypted PII, input validation, cryptography, network security, SQL injection and so on! Basically, the goal is to translate MASVS (and its sister project MSTG) into automated BDD security tests and give pentesters more time to focus on "crazy stuff"
