Agent Trust Boundaries in Security Workflows
A defensive review structure for deciding what an AI-assisted security workflow should never be trusted to do autonomously.
Research
A public archive of research notes, publication-ready articles, and defensive analysis.
A defensive review structure for deciding what an AI-assisted security workflow should never be trusted to do autonomously.
A practical set of technical and behavioral signals defenders can use to triage suspicious domains and impersonation sites.
How small device enclaves can be monitored for suspicious communication without turning the network into a full-time forensics project.
A privacy-focused review of how commercially available datasets can amplify external threat, targeting, and exposure risk.