I hold a PhD in Information Security from Royal Holloway, University of London. My research focused on IoT Security, and the title of my thesis was "Analysing and Preventing Self-Issued Voice Commands", in which I investigated to what extent commands played and captured by the same device can be a threat to the device users' security, safety and privacy. An example is CVE-2022-25809, known as the AvA Attack, which my team reported to Amazon in 2021 and was the core of my thesis.
I have been active for a few years in the consultancy field. Between 2017 and 2019, I worked for ICT Cyber Consulting, a consultancy firm in Italy. As part of my responsibilities, I used to lead the internal cybersecurity team and I supported the evaluation and the implementation of security measures for more than 60 client firms, including SMB, SME and large enterprises. Between 2023 and 2024, I worked as a freelance consultant for half-dozen client firms, mostly doing vulnerability assessments, penetration tests and training sessions on security topics.
I am now a Fixed-term Assistant Professor (RTD-A) at University of Catania. As part of my responsibilities, I teach the "Vulnerability Assessment and Penetration Testing" course, aimed at both BSc and MSc students, and I co-teach the "Internet Security" and the "Computer Security" courses, aimed at BSc and MSc students, respectively. My research activities still include IoT Security, but now also involve security of multimodal networks, as part of the PNRR FAIR project, and privacy in the automotive environment, as part of the PECS project (NGI TRUSTCHAIN).
Davide Bonaventura, Sergio Esposito, Giampaolo Bella
We extensively analyse vulnerabilities presented in our previous work and exploited against multiple devices of the Tapo Tp-Link ecosystem. We thoroughly discuss involved Tapo protocols and related exploits used to compromise the devices, using multiple modeling techniques such as flowcharts and pseudocode algorithms; we additionally provide the exploits' source code.
To appear in Elsevier Computers & Security JournalGiampaolo Bella, Gianpietro Castiglione, Sergio Esposito, Mirko Giuseppe Mangano, Mirco Marchetti, Marcello Maugeri, Mario Raciti, Salvatore Riccobene, Daniele Francesco Santamaria
The PECS project aims at revolutionising privacy management within the automotive domain, empowering users to customise privacy settings for their applications, and proposing data obfuscation solutions. We introduce the preliminary software design of PECS and five user stories to conceptualise the user interaction.
To appear in the Proceedings of the 2nd EAI International Conference on Security and Privacy in Cyber-Physical Systems and Smart Vehicles (EAI SmartSP 2024)Giampaolo Bella, Gianpietro Castiglione, Sergio Esposito, Mario Raciti, Salvatore Riccobene
We assess how well modern cars protect their drivers' data by analysing privacy policies of known car brands, finding that they are still imprecise about how brands comply with a number of GDPR articles. Hence, compliance results non-verifiable from the outside. This leads to potentially concerning scenarios in case of cyberwarfare.
To appear in Proceedings of the 2024 IEEE International Workshop on Technologies for Defense and Security (TechDefense 2024)Marcello Maugeri, Mirko Ignazio Paolo Morana, Sergio Esposito, Giampaolo Bella
We propose a Federated Learning approach, built upon the Flower Framework, that empowers users to decide if they want to participate in model training, or if they just want to make inferences without contributing to the training. We demonstrate this approach through an automotive case study that employs the EngineFaultDB dataset.
In Proceedings of the 2024 IEEE International Conference on Blockchain (Blockchain)Davide Bonaventura, Sergio Esposito, Giampaolo Bella
We show how vulnerabilities previously exploited against the Tapo L530E smart bulb can be exploited against other devices of the Tapo family, with similar impact. Additionally, we present a new attack scenario that combines the already known vulnerabilities and another network configuration in which the vulnerabilities can be exploited.
In Proceedings of the 21st International Conference on Security and Cryptography (SECRYPT 2024)Ruben Cacciato, Mario Raciti, Sergio Esposito, Giampaolo Bella
We build two relational models to describe the current privacy landscapes for Intelligent Transportation Systems (ITS) and the Internet of Vehicles (IoV), analysing differences between the two domains and highlighting key challenges for their development in the near future.
In Proceedings of the 18th International Conference on Availability, Reliability and Security (ARES '24)Sergio Esposito, Daniele Sgandurra, Giampaolo Bella
We describe VOCODES, a tailored kill chain for self-activation attacks. We show that some peculiar steps are needed to perform these attacks, compared to more classical ones. To demonstrate VOCODES' potential, we apply it to analyse the AvA attack in detail.
In Proceedings of The 6th International Workshop on Attacks and Defenses for Internet-of-Things (ADIoT 2023)Davide Bonaventura, Sergio Esposito, Giampaolo Bella
We describe four vulnerabilities found on the smart light bulb Tapo L530E by Tp-Link, by applying the PETIoT kill chain. We illustrate the impact and the execution of several attack scenarios that can be achieved by chaining said vulnerabilities, and finally we design some fixes.
In Proceedings of the 20th International Conference on Security and Cryptography (SECRYPT 2023)Giampaolo Bella, Pietro Biondi, Stefano Bognanni, Sergio Esposito
PETIoT is a tailored kill chain for penetration testing in IoT environments, with particular emphasis on traffic analysis. We illustrate all PETIoT steps and then we analyse an attack that is already known in the literature as an example, to show PETIoT's potential.
In Elsevier Internet of Things Journal, Volume 22Sergio Esposito, Daniele Sgandurra, Giampaolo Bella
We propose a taxonomy of security levels to apply to voice-controllable devices, and a countermeasure based on Twin Neural Networks to protect such devices from self-activation attacks such as AvA.
In Proceedings of the IEEE 8th European Symposium on Security and Privacy (EuroS&P 2023)Sergio Esposito, Daniele Sgandurra, Giampaolo Bella
Alexa versus Alexa (AvA) is an attack that leverages audio files containing voice commands to gain remote control of Amazon Echo devices that play them. The attack was rated as Critical by the NIST, with a CVSS severity score of 9.8 out of 10.
In Proceedings of the 2022 ACM on Asia Conference on Computer and Communications Security (ASIA CCS '22)Università degli Studi di Catania
October 2024 – Current
This course is part of both the L-31 Bachelor's Degree Course and the LM-18 Master's Degree Course in Computer Science. I taught the following course topics and designed practical lab activities for some of them:Università degli Studi di Catania
April 2024 – Current
This course is part of the L-31 Bachelor's Degree Course in Computer Science. I taught the following course topics and designed practical lab activities for some of them:Università degli Studi di Catania
February 2020 – Current
I am a tutor for the Cyberchallenge.IT (CCIT) project, preparing university and high-school students for a national hacking competition. Since the 2024 edition, I am also managing the UniCT team activities within the CCIT project.Università degli Studi di Catania
October 2021 – January 2024
I designed practical lab activities on course topics:Royal Holloway, University of London
February 2020 – July 2021
I marked coursework and I designed lab sheets on course topics:Università degli Studi di Catania
April 2019 – June 2020
I explained in detail some topics that were requested by students, during lessons and practical labs. Some relevant topics:Università degli Studi di Catania
May 2017
I was asked to give a lecture during the Internet Security course, which is part of the L-31 Bachelor's Degree Course in Computer Science. Topics covered, along with related lab activities:May 2023 - March 2024
Although in the past I used to lead GDPR compliance activities for my clients, offensive security and teaching is where I thrive.ICT Cyber Consulting
December 2018 – October 2019
ICT Legal Consulting
June 2018 – December 2018
ICT Legal Consulting
August 2017 – June 2018