Forty-two cataloged items across seven governance architectures, twelve hardware reference designs, seven software systems, twelve standalone simulations, and the publication record. Each card states what the item is, what evidence supports it, what its status label means, and what a reviewer should know before opening it. The complete 29-page technical catalog (revision 3.2, July 28, 2026) is available as a PDF, and the full index as CSV.
This catalog presents public-source, self-authored research artifacts. Evaluation is simulation-based unless explicitly stated otherwise. Hardware references are design specifications unless physical testing is specifically identified. The catalog does not represent official government endorsement, sponsorship, or affiliation.
Catalog status labels distinguish between operational physical hardware, browser-based simulations, formal models, reference designs, and future validation plans. A listed reference design should not be read as a certified product or fielded system.
Continuous sensor-trust scalar in [0,1] via weighted Dempster-Shafer fusion across internal consistency, cross-sensor agreement, temporal stability, and physical plausibility.
U.S. Provisional 64/002,453 · DOI 10.5281/zenodo.18936251
FORMAL MODEL + BROWSER-BASED SIMULATIONSingle-author research; simulation and formal evidence only; not deployed.
Graded four-level authority (A3-A0) with asymmetric hysteresis; TLA+ model-checked across 23,748 distinct reachable states at depth 9 (26,397,356 states generated); the result describes the discrete authority automaton under the assumption that instantaneous authority equals its target and does not cover continuous behaviour between decision instants; of 8 stated properties, 5 invariants and 1 liveness held, 2 vacuous at this bound (one upgrade-path, one trust-authority consistency).
U.S. Provisional 63/999,105 · DOI 10.5281/zenodo.18861653
FORMAL MODEL + BROWSER-BASED SIMULATIONSingle-author research; simulation and formal evidence only; not deployed.
Deception prior P(adversarial) that adjusts authority downward when inputs are likely manipulated; separates honest confusion from hostile attack.
DOI 10.5281/zenodo.19043924
BROWSER-BASED SIMULATIONSingle-author research; simulation and formal evidence only; not deployed.
Byzantine-resilient consensus (3f+1) with CUSUM anomaly detection; a single compromised node cannot determine consensus alone.
DOI 10.5281/zenodo.19015517 · MIT license
FORMAL SPECIFICATION + SIMULATIONSingle-author research; simulation and formal evidence only; not deployed.
Consequence-scaled mandatory deliberation window before irreversible action; five-state circuit breaker; fail-safe default on timeout is abort.
U.S. Provisional 64/005,607 · DOI 10.5281/zenodo.19015618
BROWSER-BASED SIMULATIONSingle-author research; simulation and formal evidence only; not deployed.
Deterministic GREP-phase recovery to defined safe states with a terminal non-compensatory policy gate; exercised by a self-run 10-million-iteration enumeration (site claim; not independently reproduced).
U.S. Provisional 64/000,170 · DOI 10.5281/zenodo.18917790
FORMAL ENUMERATION + SIMULATIONSingle-author research; simulation and formal evidence only; not deployed.
Quantifies decision-time compression and cross-domain escalation cascades in AI-enabled command and control; 600 Monte Carlo runs.
SSRN ID 6176802 (working paper)
BROWSER-BASED SIMULATIONSingle-author research; simulation and formal evidence only; not deployed.
Low-cost reference testbed for authority-governed autonomy; Raspberry Pi 5 + ESP32 running SATA-HMAA-CARA; seven fault scenarios demonstrated in simulation.
DOI 10.5281/zenodo.19143190 · ~$484 BOM
DESIGN SPECIFICATION; PHYSICAL VALIDATION PENDINGReference design; physical build not yet documented; results to date are simulation outcomes.
Trust-governed flight autonomy; Cube Orange+ with Jetson Orin NX companion and MAVLink HIL bridge; five adversarial scenarios demonstrated in simulation.
DOI 10.5281/zenodo.19128769 · ~$4,200 BOM
DESIGN SPECIFICATION; PHYSICAL VALIDATION PENDINGReference design; physical build not yet documented; no flight testing performed.
Reference hardware instantiation of the full nine-module pipeline with a hardwired normally-open effector interlock; 72 components.
DOI 10.5281/zenodo.19177472 · ~$139K BOM
REFERENCE DESIGN ONLY; PHYSICAL VALIDATION PENDINGHighest-cost platform; not built; MIL-STD-810G is a design target.
Authority-governed drive-by-wire safety architecture; three-leg redundant fail-safe relay; twelve attack scenarios demonstrated in simulation.
DOI 10.5281/zenodo.19232130 · ~$16.3K BOM
SIMULATION-BASED EVALUATION; PHYSICAL VALIDATION PENDINGISO 26262 ASIL-D is a target pathway, not a certification.
Hydroacoustic sonar, magnetic anomaly detection, AIS spoofing detection; sea-state authority damping; 84 components, IP68 enclosure target.
DOI 10.5281/zenodo.19246785 · ~$43K BOM
REFERENCE DESIGN ONLY; PHYSICAL VALIDATION PENDINGDesign complete; not built.
Protection node for power-grid, water, and pipeline operations; IEC 61850 GOOSE, Modbus, PROFINET; SIL-3 safety relay; 92 components.
DOI 10.5281/zenodo.19277887 · ~$11.6K BOM
REFERENCE DESIGN ONLY; PHYSICAL VALIDATION PENDINGSIL 3, NERC CIP, FIPS 140-2 are design targets.
Radiation-tolerant onboard authority under signal delay; RTG4 FPGA hot redundancy; 91 components; 15-document engineering package.
DOI 10.5281/zenodo.20183269 · ~$505K BOM
DESIGN SPECIFICATION ONLYEarliest-maturity platform; V&V specified but not executed.
Passive detection-to-authority governance across radar, RF, EO/IR, Remote ID, LIDAR; federal-to-SLTT handoff; ~75% reuse from BLADE-EDGE.
DOI 10.5281/zenodo.20299604 · ~$43.5K BOM
REFERENCE DESIGN ONLY; SIMULATION-BASED EVALUATIONAligned with EO 14305 and FY26 NDAA; alignment is a research mapping.
Tamper-evident hardware root of trust; non-exportable ECDSA keys in CC EAL6+ secure element; verified by adversarial emulator, 275/275 checks.
DOI 10.5281/zenodo.20299821 · ~$199 BOM
EMULATOR-VERIFIED DESIGN; PHYSICAL VALIDATION PENDINGSilicon not built; emulator evidence only.
BFT two-phase consensus across N = 10/50/500 agents; safe-halt-by-default under denied RF; TLA+ model-checked (5 safety and 3 liveness properties held in the analyzed model). Governs decision authority and audit, not weapons.
DOI 10.5281/zenodo.20351198 · ~$1,333/node
SIMULATION-BASED EVALUATION; PHYSICAL VALIDATION PENDINGTestbed self-assessed at TRL 2; consensus results are simulation outcomes.
Fail-closed, bump-in-the-wire appliance adjudicating each cross-boundary command across four OT regimes; malformed input fails closed.
DOI 10.5281/zenodo.20342067 · 1U appliance
REFERENCE DESIGN ONLY; SIMULATION-BASED EVALUATIONNot deployed in any facility.
Authority arbitration routing each transaction to clearance, review, confirmation, or hold; retrospective swarm-review for low-and-slow activity.
DOI 10.5281/zenodo.20374692 · ~$9,228 BOM
SIMULATION-BASED EVALUATION (SYNTHETIC DATA ONLY)Synthetic data only; not deployed in any institution.
Authority-lifecycle shim wrapping any LLM-based agent runtime; T3-T0 gating, per-tool HKDF tokens, spawn quorum, injection-collapsing input trust.
15-section specification + working simulator · hardware companion DOI 10.5281/zenodo.20299821
REFERENCE ARCHITECTURE; BROWSER-BASED DEMONSTRATIONSingle-author research; not deployed in any production environment.
An assurance gate a system must clear before it is permitted to act; mapped by the author to pre-deployment assurance themes in the March 2026 White House cyber strategy (President Trump’s Cyber Strategy for America).
Specification + governor console
REFERENCE ARCHITECTURE; BROWSER-BASED DEMONSTRATIONSingle-author research; not deployed in any production environment.
Gates AI actions across the IT/OT boundary; authorizes deterministic safety logic and does not make the safety decision itself.
Specification + governor console
REFERENCE ARCHITECTURE; BROWSER-BASED DEMONSTRATIONSingle-author research; not deployed in any production environment.
Constrains what a defensive cyber-reasoning agent may do and when; treats the agent as a black box; governance only, no offensive function.
Specification + governor console
REFERENCE ARCHITECTURE; BROWSER-BASED DEMONSTRATIONSingle-author research; not deployed in any production environment.
Gates autonomous spacecraft actions under intermittent ground contact; companion to BLADE-SPACE.
Specification + governor console
REFERENCE ARCHITECTURE; BROWSER-BASED DEMONSTRATIONSingle-author research; not deployed in any production environment.
Governs what an AI under evaluation is permitted to do inside a test environment; aligned with NDAA Sections 1533 and 1534.
Specification + governor console
REFERENCE ARCHITECTURE; BROWSER-BASED DEMONSTRATIONSingle-author research; not deployed in any production environment.
Docker-verified, synthetic-data MVP prepared for structured external technical review: agent authority-tier enforcement, human approval gating with RBAC, signed audit evidence, reviewer packet export, latency and load testing, and ledger tamper detection across clean, forged, and truncated records.
Container test suite 83 passed, 1 skipped · Python 3.11 install PASS · Docker Desktop (Mac) build, demo, and pytest PASS · QA10 final artifact after eleven AI-assisted review rounds
SYNTHETIC-DATA MVP; PREPARED FOR STRUCTURED EXTERNAL REVIEWNo operational validation or agency endorsement is claimed. Public GitHub release pending counsel review; source access is invited-review only.
Six scenarios, 600 Monte Carlo runs, model-checked property analysis, Merkle provenance; models cross-domain escalation cascades.
Client-side, seeded PRNG; bit-exact reproducible within this packaged browser build (not independently reproduced) · frameworks: ERAM · SATA · FLAME · CARA
BROWSER-BASED DEMONSTRATIONDemonstrates governance logic under seeded scenarios; does not prove field performance, hardware timing, or resistance to a live adversary.
Full seven-framework pipeline with PBFT consensus and a Merkle audit trail in a theater-level scenario.
Client-side, seeded PRNG; bit-exact reproducible within this packaged browser build (not independently reproduced) · frameworks: All seven frameworks
BROWSER-BASED DEMONSTRATIONDemonstrates governance logic under seeded scenarios; does not prove field performance, hardware timing, or resistance to a live adversary.
Air, maritime, and ground lanes with Byzantine fault isolation.
Client-side, seeded PRNG; bit-exact reproducible within this packaged browser build (not independently reproduced) · frameworks: HMAA · CARA · MAIVA
BROWSER-BASED DEMONSTRATIONDemonstrates governance logic under seeded scenarios; does not prove field performance, hardware timing, or resistance to a live adversary.
Authority-conflict interdiction via a common operating picture.
Client-side, seeded PRNG; bit-exact reproducible within this packaged browser build (not independently reproduced) · frameworks: HMAA · CARA · SATA · ADARA
BROWSER-BASED DEMONSTRATIONDemonstrates governance logic under seeded scenarios; does not prove field performance, hardware timing, or resistance to a live adversary.
Distributed Web Worker nodes, common operating picture, analytics, Merkle audit.
Client-side, seeded PRNG; bit-exact reproducible within this packaged browser build (not independently reproduced) · frameworks: All seven frameworks
BROWSER-BASED DEMONSTRATIONDemonstrates governance logic under seeded scenarios; does not prove field performance, hardware timing, or resistance to a live adversary.
Actor-model mesh with PBFT consensus and an emergent interlock.
Client-side, seeded PRNG; bit-exact reproducible within this packaged browser build (not independently reproduced) · frameworks: MAIVA · CARA
BROWSER-BASED DEMONSTRATIONDemonstrates governance logic under seeded scenarios; does not prove field performance, hardware timing, or resistance to a live adversary.
Zero-dependency kinematic engine with flight-termination authority.
Client-side, seeded PRNG; bit-exact reproducible within this packaged browser build (not independently reproduced) · frameworks: HMAA · CARA
BROWSER-BASED DEMONSTRATIONDemonstrates governance logic under seeded scenarios; does not prove field performance, hardware timing, or resistance to a live adversary.
Run-time-assurance authority gate on the ASTM F3269 and Simplex reference model; an independent invariant monitor admits, holds, or reverts each proposed authorization before it is applied.
Client-side, seeded PRNG; SHA-256 hash-chained decision ledger, replay determinism, in-browser V&V · five stress scenarios · synthetic data only
BROWSER-BASED DEMONSTRATIONStandalone research console; governance and assurance only; no vehicle dynamics, actuation, or live system; self-assessed TRL 2-4.
Engagement-authority governance gate; a fail-closed invariant monitor adjudicates each proposed action to authorize, clamp, downgrade, deny, or hold for human before any authority is granted.
Client-side, seeded PRNG; SHA-256 hash-chained decision ledger, replay determinism, in-browser verification · five seeded scenarios · synthetic data only
BROWSER-BASED DEMONSTRATIONStandalone research console; governance and assurance only; no targeting, no kinetic effects, no live system; self-assessed TRL 2-4.
Safe-state authority governance gate; a fail-safe governor adjudicates safing commands (entry, hold, deepen, and a re-arm gated exit) across the safe-state lifecycle, and the safe-state floor is never applied as an authority grant.
Client-side, deterministic; SHA-256 hash-chained decision ledger, replay against five published reference roots, in-browser verification · 1061 checks pass · formal authority model check, nine properties hold · synthetic data only
BROWSER-BASED DEMONSTRATIONStandalone research console; governance and assurance only; no dynamics, no actuation, no targeting; demonstrated in simulation; no formal TRL determination has been performed.
Team authority governance prototype; an independent arbiter over an untrusted allocator maintains exactly one eligible authenticated holder per responsibility, never zero and never two, with bounded reversible transfers, a safe-hold fallback, and an allowlisted supervisor override.
Client-side, deterministic; SHA-256 hash-chained decision ledger, deterministic replay from seed 20260727 reproducing the published reference ledger root, in-browser scenario suite · 20 of 20 scripted scenarios contained · boot self-test, engine encapsulation · synthetic data only
HARDENED INTERACTIVE PROTOTYPEStandalone research console, hardened interactive prototype v0.16, not a reproducible signed release; governance and assurance only; no dynamics, no actuation, no targeting; demonstrated in simulation; no formal TRL determination has been performed.
A deterministic synthetic console in which the critical authorization can be held only by two distinct authenticated humans, a machine is never eligible, and the failure safe state is NO-GO.
Client-side, deterministic; SHA-256 hash-chained decision ledger, 27 of 27 bundled deterministic scenarios contained, all six audits pass; seeded sweep, 250 seeds, 25,000 ticks, zero audit failures · synthetic data only
HARDENED STANDALONE PROTOTYPEStandalone research console, hardened standalone prototype v0.13, not a reproducible signed release; governance and assurance only; no dynamics, no actuation, no targeting; demonstrated in simulation; no formal TRL determination has been performed.
Zenodo deposits with registered DOIs and SSRN working papers; a record-level publication registry is pending, so no single total is published until the site catalog and the July 2026 study inventory are reconciled. Public disclosures with permanent identifiers; not peer-reviewed venues. Source-code licenses are stated per record.
Zenodo + SSRN + ORCID 0009-0001-8573-1667
PUBLISHED / DISCLOSEDIndependent third-party citation is currently limited; one peer-reviewed journal article is published (see EVID-04); the corpus itself is self-archived.
HMAA 63/999,105 · CARA 64/000,170 · SATA 64/002,453 · FLAME 64/005,607 (filed March 2026, USPTO receipts); ADARA 64/110,218 · MAIVA 64/110,221 · ERAM 64/110,225 (reported filed July 13, 2026, receipts pending). Unexamined filings, not granted patents; potential later nonprovisional filing subject to applicable requirements and counsel advice.
USPTO Patent Center
PROVISIONAL; PENDING USPTO ACTIONProvisional applications reported as filed; unexamined; a later nonprovisional filing may claim benefit subject to applicable requirements and counsel advice.
The Authority Equation (3), The Authority Discipline (4), Autonomous Authority (3), published under the Authority & Architecture imprint; ISBNs via Bowker; Volume I carries LCCN 2026912260.
LCCN 2026912260 · Bowker ISBNs
PUBLISHEDBook series is technical reference material, not peer-reviewed literature.
Oktenli, B. SATA: Sensor Attestation and Trust Anchoring: A Formally Specified, Simulation-Evaluated Continuous-Trust Protocol for Autonomous-Systems Sensor Edges. Journal of Hardware and Systems Security (Springer Nature) 10, 17 (2026). Received 4 June 2026; accepted 14 July 2026; published 21 July 2026.
DOI 10.1007/s41635-026-00190-4 · Springer Nature
PUBLISHED · PEER-REVIEWEDFirst peer-reviewed journal article of the program; the published protocol corresponds to the SATA architecture (AUTHREX-01).
| EVIDENCE TYPE | WHAT IT IS | SUPPORTS | DOES NOT PROVE |
|---|---|---|---|
| Zenodo deposits (18) | Architecture and platform specifications, BOMs, code | Public disclosure, priority, reproducibility | Field performance, certification |
| SSRN working papers (15) | Frameworks and policy analysis incl. ERAM | Public disclosure with permanent IDs | Peer review |
| Peer-reviewed journal article (1) | SATA protocol in the Journal of Hardware and Systems Security (Springer Nature, 2026) | Independent editorial peer review of the published protocol | Field performance, certification |
| Reported TLA+ model checking (HMAA) | 23,748 distinct reachable states at depth 9 (26,397,356 states generated); the result describes the discrete authority automaton under the assumption that instantaneous authority equals its target and does not cover continuous behaviour between decision instants; of 8 stated properties, 5 invariants and 1 liveness held | Model-checked properties of the authority state machine | Correctness of other frameworks; hardware timing |
| Provisional patent applications (7) | SATA, HMAA, CARA, FLAME, ADARA, MAIVA, ERAM filings | Priority dates and USPTO receipts | Granted patent rights |
| Browser simulations (22) | Seeded, client-side, reproducible | Governance logic behavior under scenarios | Resistance to a live adversary; hardware behavior |
| Reference designs (12) | BOM-level hardware packages | Buildability analysis and cost baselines | Built hardware; measured interlock behavior |
| ITEM | RESEARCH STATUS | EVIDENCE | LIMITATION | NEXT VALIDATION STEP |
|---|---|---|---|---|
| AUTHREX pipeline (7 architectures) | Formal model + browser-based simulation | DOIs, TLA+ (HMAA), seeded sims | Simulation and formal evidence only | FPGA bitstream + hardware-in-the-loop |
| BLADE hardware platforms (12) | Reference design / design specification | Zenodo deposits, full BOMs | Not built; no physical testing | Build Rover and UAV testbeds first |
| Software reference architectures (6) | Reference architecture | Specifications + governor consoles | Not deployed in production | Independent red-team evaluation |
| Simulation portfolio (22) | Browser-based demonstration | Seeded, bit-exact reproducible within this packaged browser build (not independently reproduced) runs | Not validated against operational data | Sensitivity analysis; peer review |
| Publications and IP | Published / disclosed | DOI, SSRN, USPTO receipts, LCCN | Self-archived corpus; one peer-reviewed journal article (Springer Nature, 2026) | Additional peer-reviewed submissions; potential later nonprovisional filing (subject to requirements and counsel advice) |
Start with the Executive Overview and System Boundary in the PDF catalog, then BLADE-EDGE and BLADE-CUAS cards above, then run the APEX JADC2 Theater simulation. The program white paper states the proposed physical validation campaign.
Start with the HMAA deposit (reported TLA+ model checking, 23,748 distinct reachable states at depth 9 (26,397,356 states generated); the result describes the discrete authority automaton under the assumption that instantaneous authority equals its target and does not cover continuous behaviour between decision instants) and the Rover Testbed package, then the Risk and Limitation Assessment in the PDF. Every simulation is seeded for bit-exact reproduction within this packaged browser build; verify one independently before reading further.
Start with AUTHREX-AGENT and BLADE-AGENT-HSM, then the AUTHREX-AGENT-CYBER card, then run the AGENT governor console. Alignment references: CISA/NSA/Five Eyes agentic-AI guidance and FY26 NDAA Sections 1513 and 6601.
Start with BLADE-SPACE (15-document engineering package) and AUTHREX-SPACECYBER, then the orbital scenario in the Operational Domains section of the main site. Note: BLADE-SPACE V&V is specified but not executed.
Start with BLADE-MARITIME (sea-state authority damping, acoustic-delay-aware consensus), then the MDO Digital Twin simulation's maritime lane and the maritime mission panel on the main site.
Machine-readable project data is embedded in this page as JSON (view source, #catData) and mirrored in the CSV index. Missing or unverifiable links are marked Needs verification in the source data.