Incalmo: An Autonomous LLM-assisted System for Red Teaming Multi-Host Networks
B. Singer, K. Lucas, L. Adiga, M. Jain, L. Bauer, and V. Sekar. IEEE S&P '26. [BibTeX]Get details or download paper
Publications
Technical Reports & Industry Writing
Not peer-reviewed; published on Anthropic’s Frontier Red Team blog, where we post our work measuring, projecting, and improving frontier LLM capabilities.
Measuring LLMs' Impact on N-day Exploits 2026
How well frontier models can develop exploits for publicly disclosed but unpatched vulnerabilities.
Measuring LLMs' Ability to Develop Exploits 2026
Benchmarks (ExploitBench, ExploitGym, SCONE-bench) for measuring how well frontier models build end-to-end software and smart-contract exploits.
Assessing Claude Mythos Preview's Cybersecurity Capabilities 2026
Evaluating a frontier model's ability to autonomously discover and exploit zero-day and N-day vulnerabilities across operating systems, browsers, and libraries.
Partnering with Mozilla to Improve Firefox's Security 2026
A collaboration with Mozilla in which Claude Opus 4.6 found 22 vulnerabilities in Firefox.
Reverse Engineering Claude's CVE-2026-2796 Exploit 2026
A technical deep dive into an exploit Claude developed for a Firefox vulnerability.
AI Models on Realistic Cyber Ranges 2026
Claude succeeds at multistage network attacks using only standard tools.
Claude in Cyber Competitions 2025
Evaluating Claude's performance in CTF competitions (the subject of my DEF CON talk).
Cyber Toolkits 2025
A collaboration with Brian Singer on measuring LLMs' network attack capabilities (related to the Incalmo paper below).
Peer-Reviewed Publications
Get details or download paper
Training Robust ML-based Raw-Binary Malware Detectors in Hours, not Months
K. Lucas, W. Lin, L. Bauer, M. K. Reiter, and M. Sharif. CCS '24. [BibTeX]Get details or download paper
Group-based Robustness: A General Framework for Customized Robustness in the Real World
W. Lin, K. Lucas, N. Eyal, L. Bauer, M. K. Reiter, and M. Sharif. NDSS '24. [BibTeX]Get details
RS-Del: Edit Distance Robustness Certificates for Sequence Classifiers via Randomized Deletion
Z. Huang, N. G Marchant, K. Lucas, L. Bauer, O. Ohrimenko, B. IP Rubinstein. NeurIPS '23. [BibTeX]Get details
Adversarial Training for Raw-Binary Malware Classifiers
K. Lucas, S. Pai, W. Lin, L. Bauer, M. K. Reiter, and M. Sharif. USENIX Security '23. [BibTeX]Get details or download paper
Perspectives from a Comprehensive Evaluation of Reconstruction-based Anomaly Detection in Industrial Control Systems
C. Fung, S. Srinarasi, K. Lucas, H. B. Phee, L. Bauer. ESORICS'22. [BibTeX]Get details or download paper
Constrained Gradient Descent: A Powerful and Principled Evasion Attack Against Neural Networks
W. Lin, K. Lucas, L. Bauer, M. K. Reiter, M. Sharif. ICML'22. [BibTeX]Get details or download paper
Any-Play: an Intrinsic Augmentation for Zero-Shot Coordination
K. Lucas, R. Allen. AAMAS'22. [BibTeX]Get details or download paper
Malware Makeover: Breaking ML-based Static Analysis by Modifying Executable Bytes
K. Lucas, M. Sharif, L. Bauer, M. K. Reiter, and S. Shintre. AsiaCCS'21. [BibTeX]Get details or download paper
Creating xBD: A Dataset for Assessing Building Damage from Satellite Imagery
R. Gupta, B. Goodman, N. Patel, R. Hosfelt, S. Sajeev, E. Heim, J. Doshi, K. Lucas, H. Choset, M. Gaston. CVPR Workshops '19. [BibTeX]Get details