Private applied-research laboratory building defensive systems at the intersection of cognitive science, adversarial intelligence, and information integrity. Operating where precision matters and ambiguity is the threat.
Esothel Labs is a private research and engineering company specializing in cognitive security, influence-operation defense, and adversarial AI systems. We build tools and frameworks that detect, classify, and neutralize deceptive patterns in language, media, and autonomous agent behavior.
Our work sits at the boundary between what is said and what is meant — where manipulation hides in structure, not content. We formalize the informal. We instrument the invisible.
Each project addresses a distinct vector in the cognitive-security problem space. From formal threat specification to real-time detection infrastructure.
Multi-layer detection framework for behavioral manipulation, influence operations, and adversarial content. Includes formal threat specification via HPL (Hostile Pattern Language), a six-layer detection pipeline, and APL (Adversarial Probing Language) for symmetric defense testing. Python and Node.js implementations. Designed for humanitarian and defensive application.
Research program mapping the structural grammar of influence — recurring patterns in rhetoric, framing, and narrative construction that transcend language and platform. Feeds directly into M.I.N.D.'s detection taxonomy and HPL primitive definitions.
Extension of M.I.N.D. principles to autonomous AI agents. Real-time detection of adversarial prompt injection, behavioral drift, and goal manipulation in agent pipelines. Protecting the next generation of AI systems from the threats they weren't designed to see.
Curated, machine-readable feed of active influence operation signatures, adversarial patterns, and cognitive threat indicators. Designed for integration with defensive platforms, SOC workflows, and academic research pipelines.
For research inquiries, partnership discussions, or responsible disclosure.