Research

Credibility is not claimed. It is published.

Field notes from the seam where code, intelligence, and narrative meet—written for practitioners, not for search engines.

Featured researchFirst paper

AI application security

The Unseen Attack Surface

Why the most dangerous LLM-application vulnerabilities live in the model's interpretation of language—where no conventional scanner can follow.

Open the research index

Why we publish

A security firm earns trust in the open.

Capability should be examined, not inferred from presentation. Where we can prove something, we publish the method and the evidence. Where the field is still learning, we say so plainly.

Research is how the practice stays honest: every claim must survive contact with another capable mind.

Research streams

Work organized around what defenders can use.

Three publication streams connect original technical work, regulatory interpretation, and current threat context.

Original findings

Technical research & disclosures

Original analysis and responsibly disclosed vulnerabilities, including why consequential AI-application weaknesses remain invisible to conventional scanners.

Browse the index

Operating guidance

Compliance guides

Defender-focused readings of DORA, NIS2, and the EU AI Act for institutions that must translate obligation into operating practice.

Browse the index

Frontier context

Threat & AI-security briefings

Recurring analysis of emerging threats at the frontier of artificial intelligence, written for practitioners rather than search engines.

Browse the index

Publication standard

From observation to useful record.

Every publication keeps the path from original signal to defensible conclusion visible.

Observe

Record the signal without forcing it into a familiar category.

Prove

Reproduce the behavior and separate evidence from interpretation.

Disclose

Give the responsible owner time and context to act.

Publish

Share what materially improves the field.

Current index

Research for the moment before a pattern becomes obvious.

The Briefing Hub keeps research, operational notes, and defender guidance in one legible publication system.

Visit the Briefing Hub

Technical researchThe Unseen Attack SurfaceAI application security Regulatory guideDORA, NIS2, and operational evidenceEuropean resilience Threat briefingSignals at the AI-security frontierEmerging behavior

Research-led by design

Examine the work before you trust the claim.

Browse the research index or begin a confidential conversation about your environment.

Get a demo About the practice