Quandary Peak Adds CogniCrypt Malware Detection to M&A Due Diligence

Quandary Peak Research has added CogniCrypt, a framework for detecting malware generated by AI, to its technical due diligence process for M&A and technology investments, the company announced in a press release.

CogniCrypt analyzes unfamiliar software behavior, prioritizes execution paths, and classifies possible malware using machine learning. The framework combines concolic execution, path prioritization guided by large language models, and transformer based classification. Its experimental evaluation reported 97.5% detection accuracy on malware samples generated by AI.

Quandary Peak said its due diligence process identifies critical systems, data, software dependencies, and operational requirements before reviewing source code, system architecture, security controls, known incidents, and threat exposure. Technical findings are evaluated by scope, severity, likelihood, and possible business effects.

The CogniCrypt framework is described in a paper titled CogniCrypt: Synergistic Directed Execution and LLM Driven Analysis for Zero Day AI Generated Malware Detection. The paper was coauthored by George Edwards and Mahdi Eslamimehr and accepted by the 12th International Conference on Cryptography and Information Security, CRIS 2026, in Zurich.