Tamandua builds on research, open standards, and security community tooling.
Tamandua's ML malware scoring layer is inspired by and implements concepts from Malware-SMELL, a zero-shot learning strategy for malware detection using binary visualization, deep metric learning, and similarity-space reasoning.
Reference:
- Pedro H. Barros, Eduarda T. C. Chagas, Leonardo B. Oliveira, Fabiane Queiroz, Heitor S. Ramos. "Malware-SMELL: A zero-shot learning strategy for detecting zero-day vulnerabilities." Computers & Security, Volume 120, 2022, 102785. https://doi.org/10.1016/j.cose.2022.102785
The Tamandua team is grateful for the technical support and guidance from the Malware-SMELL creator/research team as we adapt the approach to an operational EDR/XDR setting.
Tamandua also integrates or interoperates with community security formats and tooling such as YARA, Sigma, MITRE ATT&CK, STIX/TAXII, and OSINT threat-intelligence feeds. Each integration should retain the relevant upstream license and attribution when distributed.