About LRRK

Evidence for consequential decisions.

LrrK studies the security of connected and autonomous systems, and the software, communications, and service dependencies that shape their behavior. Our working model connects scoped evidence, validation, human review, and a record that can be reassessed as systems change.

LRRK LLC offers independent cyber-physical assurance for systems that sense, move, connect and act. It is for the people who must decide whether to deploy, update, operate or replace one of those systems and need inspectable evidence to do it. The Atlas plates, gates and lab exhibits on this site are demonstrations; requests begin with a scoped question reviewed by a person. The November 1 standard v0.1 operating model is proposed, not evidence of proven operating effectiveness.

List your device’s properties in the MITRE EMB3D™ mapper to see which threats and mitigations are worth asking about. The results are planning candidates, not findings. Current engagements focus on consumer robotics; suitability and any testing are confirmed during scope review.

Attributes →

From decision to practical handoff

Decision → bounded question → scope and owner authority → source/test → claim → independent human review → delivered decision record → practical handoff → observed outcome and learning.

Evidence Brief is the entry model: one bounded question, explicit source/method boundary, reviewed conclusion and reusable evidence packet. Assurance Desk is recurring human-governed scoped review or cadence after scope review. Current engagements focus on consumer robotics. Controlled testing requires an agreed scope, owner authority, suitable access and capability, and recovery and stop conditions.

Evidence Brief →Assurance Desk →

Principles the archive is built on

Traceable review: The proposed method preserves evidence locators, review decisions and reasons for revising a conclusion.
Declared claim types: Distinguish researched mechanisms, controlled demonstrations, operational incidents and scoped findings; separate observation from inference.
Scoped forever: A finding carries its configuration, period and method wherever it is quoted.
Capability honesty: Designed workflows are labelled as designed. Nothing implies an action that did not occur.

Where to start

Pricing

Consumer robots are arriving in homes, buildings and fleets faster than anyone is documenting them. LRRK publishes a free public register and accepts human-reviewed consulting requests. Robot Purchase Check is a bounded pilot whose checkout remains closed; its stated terms are available on the pricing page.

Pricing — engagement terms →

Stated limits

The Atlas plates and lab demonstration interfaces are a designed workflow: fictional subjects, synthetic evidence, read-only. Nothing in those demonstrations is collected, hashed, signed, published, remediated or monitored, and no control in them implies that such an action occurred.

Unlike those fixtures, the evidence brief and assessment request forms, and account sign-in, record what you submit so a person can review and scope it. The research library publishes fixed papers you can download and cite.

A Passport is point-in-time and device-specific. It is not certification, approval, a warranty, or a guarantee of safety.