security

AI systems are starting to behave less like passive tools and more like autonomous attack surfaces. A technical look at prompt injection, a concrete end-to-end attack chain, a scoping of which architectures are actually at risk, and practical defensive actions for engineering teams.

Fackel is a multi-agent pentest framework where LLMs decide strategy, not hardcoded pipelines. A walkthrough of the architecture, the design decisions, and the lessons learned.

A practical deep dive into device code phishing combined with vishing targeting Microsoft Entra: how the OAuth device code flow gets abused, what to monitor, and how to mitigate.

A practical overview of modern AI agent systems: tool use, retrieval, memory, verification, multi-agent patterns, evaluation, and security.

A rigorous analysis of how probabilistic reasoning in generative models shapes security risk, failure modes, and robustness.

Argues that abstraction layers can obscure failure modes, shift risk across boundaries, and weaken assurance unless their assumptions are made explicit.

Argues that probabilistic behavior, distributional risk, and system composability invalidate core assumptions of classical threat modeling for generative AI.