IVD has two prongs: IVD-N helps protect network availability, and IVD-ACP checks whether inputs should be trusted before they can act.
IVD-N: is designed to identify coordinated traffic patterns and support limited, auditable filtering earlier in the path before attack traffic reaches the protected edge.
IVD-ACP: is designed to evaluate packages, commands, prompts, documents, artifacts, or tool calls before they become trusted or executable.
Controlled-environment validation evidence exists. TRL-7-oriented independent testing is underway with Gear Six Labs, with completion targeted for June 25, 2026. Federal lab evaluation remains pending.
IVD is not presented as production-accredited, as having completed federal-lab validation, or as production-deployed at this stage.
Controlled or relevant environment validation evidence prepared for independent review.
Scenario artifacts are organized for independent review rather than ad hoc demonstration.
Commercial and institutional discussions are intended to begin with defined scope, not open-ended claims.
Filed U.S. non-provisional utility applications cover the dual-prong control-plane architecture.
In network environments, the cost spike arrives when hostile traffic has already formed into a service-impacting object. In AI and software environments, the damage window opens when unsafe content is admitted into trusted memory, indexing, or execution paths before policy meaningfully engages.
IVD is built around the opposite sequence: identify structure early, reduce it to a limited and auditable decision, and act before the damage becomes expensive or difficult to reverse.
IVD converts systemic attack behavior into limited, auditable policy decisions. It does not claim to replace every downstream control. It is intended to reduce the number of failure paths that reach those controls in the first place.
Identify invariant patterns across traffic, artifacts, and control surfaces.
Assign explicit outcomes and mitigation decisions before unsafe actions spread.
Preserve service continuity through scoped and auditable controls.
IVD-N is designed for environments where traffic becomes most dangerous after many sources concentrate on one target. Instead of waiting for hostile traffic to saturate the victim's edge, IVD-N looks for stable traffic behavior earlier in the path and converts coordinated activity into limited mitigation decisions.
IVD-ACP addresses a different failure path: unsafe artifacts, commands, or administrative actions entering trusted workflows before policy meaningfully engages. ACP checks whether an input should be trusted before execution, indexing, retrieval eligibility, or privileged action.
Current claims are limited to controlled or relevant environment validation, frozen evidence bundles, and evaluation-ready materials for qualified reviewers.
Controlled-environment validation evidence exists; TRL-7-oriented third-party testing is underway with completion targeted for June 25, 2026. Federal lab evaluation remains pending.
IVD is positioned for service-preserving network stabilization and safer execution governance. It is not presented here as a production deployment claim, accreditation claim, or government-backed approval claim.
Unified overview of IVD-N and IVD-ACP as complementary security architectures.
Download PDFTechnical overview of the Admissibility Control Plane for pre-execution and pre-index control.
Download PDFCanonical exploit-card destination with technical exhibit previews and full-screen timelines.
View libraryTRL-6 validation scope, review posture, and evaluation boundaries.
ReviewFive minutes. Not live. See how IVD-N and IVD-ACP make limited, auditable policy decisions.
Open simulatorInvariant Vector Defense addresses two related failure paths. IVD-N protects network availability through earlier pattern-based mitigation. IVD-ACP protects execution trust by evaluating unsafe actions, artifacts, commands, and inputs before they are trusted, indexed, or executed.
