IVD-N controls coordinated malicious traffic before network convergence. IVD-ACP controls malicious or non-admissible execution influence before indexing, persistence, or action.
IVD-N: distributed evidence becomes one bounded attack object, a governed mitigation decision, and an auditable lifecycle.
IVD-ACP: artifacts, commands, retrieved content, memory writes, and tool actions receive explicit authority before they can influence execution.
Two filed U.S. non-provisional patent applications. Controlled validation across network and execution domains. Independent third-party verification and validated reproduction within defined scope.
ACP can enter through one RAG ingestion gate, agent-tool boundary, CI/CD workflow, administrative API, or privileged automation surface without requiring carrier participation or router-policy change.
Its commercial advantage is breadth. Its competitive burden is proving that deterministic admission, authority state, durable policy, and replay are more valuable than bundled prompt screening.
IVD-N requires more demanding network access, router validation, and operator trust, but addresses a high-consequence control-plane problem with a distinctive architecture and strategic relevance to carriers, critical infrastructure, and federal networks.
Its commercial burden is validation friction. Its potential advantage is deeper infrastructure defensibility once operationally proven.
Distributed attacks are larger, more adaptive, and more costly once they reach downstream chokepoints. Automation helps only when install, scope, expiry, withdrawal, rollback, and audit are governed.
AI agents, RAG systems, administrative APIs, CI/CD pipelines, and automated workflows can turn retrieved content or artifacts into privileged behavior. The control question is whether that influence should become execution-eligible at all.
In both domains, detection alone is no longer the central issue. The issue is whether harmful behavior becomes operational before a trusted control plane intervenes.
IVD-N identifies coordinated behavior across distributed observations, forms a canonical attack object, and translates that object into bounded upstream mitigation. The value is not FlowSpec alone; it is the governed chain from evidence to install, expiry, withdrawal, rollback, relief, and reconstruction.
IVD-ACP evaluates artifacts, commands, retrieved content, memory writes, and tool actions before they become execution-eligible. It assigns bounded authority states and preserves durable, replayable enforcement records outside the system seeking permission to act.
IVD-ACP has completed controlled adversarial validation against real-world autonomous-agent failure modes including credential misuse, unauthorized delegation, alternate egress, persistence, and cross-session execution influence. Review the bounded result.
| Product | Built | Controlled evidence | Current commercial path |
|---|---|---|---|
| IVD-N | Prototype control-plane system | Macro-object, policy, multi-vendor virtual-router, and mitigation-lifecycle evidence | Paid technical evaluation, design partnership, and licensing discussion |
| IVD-ACP | Prototype admission-control system | Pre-index, authority-state, execution, replay, restart, and quarantine evidence within defined scope | Paid protected-surface evaluation, design partnership, and licensing discussion |
IVD maintains a TRL-6 public baseline. Controlled evidence and independent third-party validation do not constitute production deployment, accreditation, completed federal-lab validation, physical provider-edge validation, or a blanket TRL-7/TRL-8 claim.
Directed to distributed invariant-vector detection, macro-object synthesis, bounded upstream mitigation, service-aware relief, and related control-plane coordination.
Directed to pre-index and pre-execution admissibility control, authority-state enforcement, quarantine, replay, and separation of decision authority from execution.
Patent applications are pending. Claim scope remains subject to USPTO examination; these summaries are illustrative and do not define the claims under review.
Defined-duration review with test plans, evidence packages, technical documentation, and bounded acceptance criteria.
Controlled IVD-N or IVD-ACP deployment against one protected domain, surface, workflow, or operational environment.
Potential integration into network-security, routing, cloud, AI-security, SaaS, and automation platforms.
Investment or corporate partnership to accelerate field validation, hardening, patent strategy, and go-to-market execution.
IVD is designed for direct deployment, design-partner evaluation, and potential integration into established network-security, AI-security, cloud, routing, and automation platforms. The two products share architectural discipline and evidence infrastructure while retaining separate buyers, deployment paths, and commercial models.
Physical provider-edge and real-traffic shadow-mode validation for IVD-N.
Live RAG, agent-tool, CI/CD, or privileged-automation evaluation for IVD-ACP.
High availability, multi-tenancy, integrations, operator workflows, and continued patent prosecution.
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.
ReviewSee how IVD-N and IVD-ACP make bounded, auditable policy decisions. The simulator illustrates doctrine; it is not a production environment or validation substitute.
Open simulatorIVD is engaging qualified evaluators, design partners, strategic investors, federal and critical-infrastructure stakeholders, and platform companies interested in network resilience or execution integrity.
