AI red teaming | Security engineering

PhD researcher in adversarial machine learning, focused on the AI that defends industrial and vehicle networks.

Website · LinkedIn · ORCID · Email

• Found and reported CVE-2026-91141 (COVESA vsomeip) and CVE-2026-58196 (Stacklok ToolHive)
• 12 peer-reviewed papers: 8 published, 4 to appear
• 28 merged upstream PRs, 12 of them in Microsoft PyRIT

I found both vulnerabilities and reported them through coordinated disclosure. Both are fixed upstream.

• CVE-2026-91141, COVESA vsomeip. A 16-byte SOME/IP header made an endpoint allocate 512 MiB before the payload arrived. SOME/IP carries service communication inside vehicles. CERT/CC VU#128947 · Writeup
• CVE-2026-58196, Stacklok ToolHive. A remote MCP server could make the host runtime send requests to internal addresses, outside the container sandbox. Advisory · Writeup

• Microsoft PyRIT: 12. Stop encoded-prompt echoes from counting as successful decodes; fix Unicode payload decoding and package hallucination checks.
• cantools: 10. Fix candump log parsing, SYM and KCD writers, and decoded-log plotting.
• Scapy: 2. Dissect the PROFINET PRAL_AlarmItem; correct the Modbus Report Slave ID byte count.
• opcua-asyncio: 1. Stop swapping send and receive buffer sizes in the Hello/Acknowledge exchange.

All 28 merged PRs

To appear in 2026:

• When the Analyst Scans for the Attacker: Subverting Agentic LLM Security Operations in OT Networks. SHIELD-AI @ ECML PKDD 2026, Springer CCIS
• KEYSER: Stateful Adversarial Campaign Evaluation of ML-Based Intrusion Detection in Industrial Automation. IEEE ETFA 2026
• Adversarial Robustness Evaluation of a Multi-Layer OT Intrusion Detection System via Automated C2 Red Teaming. IEEE ETFA 2026
• Agentic Incident Response in IT/OT Critical Infrastructure: Reinforcement Learning and Cross-Domain Generalization. CCSS 2026

Published:

• Characterizing Domain Shift in PROFINET Intrusion Detection: A Multi-Site Ground-Truth Evaluation. IEEE WFCS 2026
• CoAnalyst: An Agentic LLM-Based Cybersecurity Analyst for Industrial Control Systems. IEEE WFCS 2026
• An Integrated Testbed for MITRE-Mapped Attack Emulation in Industrial Control Networks. MDPI Sensors 2026
• An Operational Hybrid SIEM Framework for OT Anomaly Detection. MDPI Sensors 2026
• A SIEM-Based Framework for Multi-Layer Data Collection and Anomaly Detection in OT Networks. IEEE IDAACS 2025
• A Testbed for Cyber Attack Emulation and AI-Driven Anomaly Detection in Industrial IoT and OT Networks. IEEE IDAACS 2025
• A Comprehensive Review of Public Datasets for Machine Learning-Based Intrusion Detection in IoT and OT Networks. IEEE IDAACS 2025
• Mapping the MITRE ATT&CK Framework to Modbus Cyber Attacks in Industrial OT Networks. IEEE ICEST 2025

• tank-lab. A Modbus attack overflows a simulated water tank while the ML intrusion detector reports normal; a physics check on the flow meters catches it. Live demo
• AutoMCP. Code for CoAnalyst, the LangGraph analyst for OT/IT alert triage with human review from the IEEE WFCS 2026 paper.
• codex-repair-shop. Each episode takes one broken AI agent, scans it, fixes it and proves the fix with tests.
• agentarium. A local, read-only view of what AI agents did across projects, with the evidence behind each result.