Living Labs

Cyber Lab

We provide advanced static, dynamic, and behavioural analysis of malicious software. Our services include malware classification, identification of malware families, analysis of API calls, system activity, network communication, persistence mechanisms, and evasion techniques. We also investigate polymorphic, obfuscated, and AI-enabled malware whose behaviour may change dynamically during execution.

We design and validate experimental AI systems for threat detection, incident analysis, alert prioritisation, and defensive decision support. These solutions may use machine learning, LLM agents, behavioural analytics, and multi-agent architectures to support SOC teams and security analysts. The result is typically a proof of concept, demonstrator, or prototype tailored to a specific defensive scenario.

We combine conventional penetration-testing methodologies with artificial intelligence and automated reasoning. AI-supported tools can help identify attack paths, prioritise vulnerabilities, analyse complex findings, and adapt testing strategies to the target environment. The service can include targeted security assessments, red-team scenarios, and experimental agent-based penetration testing.

We provide expert studies, feasibility analyses, technology assessments, specialised training, and research support in AI, cybersecurity, and quantum computing. Activities can be tailored to technical teams, management, public authorities, or research organisations. The service may also include the design of research roadmaps, datasets, experimental methodologies, and collaborative R&D projects.

We assess whether selected cybersecurity problems can benefit from quantum or quantum-inspired computation. Services may include the design of hybrid quantum-classical workflows, optimisation models, quantum classifiers, and benchmarking against conventional methods. The emphasis is placed on technically realistic use cases and measurable benefits rather than purely theoretical demonstrations.

We evaluate the security, robustness, and misuse resistance of AI applications, large language models, RAG systems, and autonomous agents. Testing may include prompt injection, jailbreaks, data leakage, adversarial inputs, poisoned knowledge sources, unsafe tool use, and excessive agent permissions. The objective is to identify technical weaknesses before they can be exploited in real-world operation.