Capability Roadmap

Where we are. Where we are headed.

Six capabilities. SOVEREIGN III is operational with full receipts — F1 0.9915 on Samsung CredData, beating the academic best by 1.85pp. The remaining intelligence and authority layers are in active development against the constraints SOVEREIGN III now satisfies. Two service lines — Intelligent System and Rapid Prototyping — are generating engagements today. This is where each line stands.

SOVEREIGN III
Operational · F1 0.9915
100%
Code and credential protection scanner. F1 0.9915 full / 0.9900 5-fold CV on the Samsung CredData benchmark (332 repos). Beats the academic best (UMass GPT2-MLP, F1 0.973) by 1.85pp using three open-source pattern engines plus a Sparked proprietary scanner plus a sklearn logistic regression meta-learner — no GPU required. Currently undergoing evaluation against the SecretBenchSecretBench Academic BenchmarkPeer-reviewed credential-detection benchmark from North Carolina State University, published via the NSF Public Access Repository. 97,479 secrets across 49 programming languages, 15,084 manually labeled true secrets across 818 GitHub repositories.Click for the official source → academic benchmark from North Carolina State University. See full receipts →
DLINE
Theoretical · In Development
50%
Constitutional authority substrate being developed. Designed for a four-phase adversarial examination on every consequential decision, with a named human authority on every authorization and a tamper-evident audit log mapped to NIST 800-53NIST Special Publication 800-53Security and Privacy Controls for Information Systems and Organizations. Includes AU-9 (Audit Information Protection), AU-10 (Non-repudiation), and RA-5 (Vulnerability Monitoring).Click for the official source →, DoD Directive 3000.09DoD Directive 3000.09 — Autonomy in Weapon SystemsU.S. Department of Defense directive governing autonomous and semi-autonomous weapon systems. Requires appropriate human judgment over the use of force and rigorous testing, evaluation, and review of AI-enabled systems before fielding.Click for the official source →, and EU AI Act Article 14European Union Artificial Intelligence ActEU regulation establishing risk-based obligations for AI systems. Article 14 requires effective human oversight of high-risk AI systems, including the ability to intervene in or override AI-driven decisions.Click for the official source →. Provisional patent filing pending before architectural detail is published.
MIMIC
Theoretical · In Development
50%
Sovereign intelligence layer being developed. Designed to transform into a specialized model that mimics each customer's operational needs. Built on foundation models with established federal deployment history. The model comes to the data; data does not leave the operational boundary. Architecture under active SBIR Phase I framing.
SENTINEL
Theoretical · In Development
50%
Twelve-phase physical security platform being developed. Designed to integrate with SOVEREIGN III findings and route through the DLINE authority substrate for cross-capability handoffs. Build sequence is intentional — SENTINEL inherits from the primitives the other two capabilities define.
Intelligent System · SBIS / FAIS
Operational · Customer Delivery
In Production
One system, two mission profiles. Deployed as the Small Business Intelligent System (SBIS) for small business owners, and as the Federal Agency Intelligent System (FAIS) for federal agency program offices serving those same owners. Operational in customer delivery under NDA.
Rapid Prototyping & MVP Development
Live · Service Line
Live
Working interactive demos and minimum viable products for federal program offices, accelerator programs, university entrepreneurship initiatives, and prime contractor innovation arms. Live and generating engagements today, with a defined path from demo to production build.
Continue
Capabilities
See every capability in detail.
Continue