Each simulation is a browser-based console for the corresponding architecture or technical report. The consoles demonstrate the mechanisms described in the published specifications, with real-time computation and cryptographic operations. Known deviations include FLAME's signing step, which is a placeholder, and MAIVA's quorum check, which has an open defect; the modeling status of each architecture is recorded on the Formal Coverage Status page.
Click any tile to jump to that simulation card below. All 39 simulations open in a new tab: 19 hosted on this site from each card's "Launch" button, and 20 on AUTHREX Systems.
Browser-based JavaScript consoles that demonstrate the mechanisms described in the published specifications (known deviations are noted above). Each card opens the standalone simulation in a new tab.
Browser implementation of the real-time authority computation engine. Includes the four-tier HMAS authority spectrum, live calculator with parameter sliders, escalation sweep analysis, EW hysteresis modeling, Monte Carlo tier distribution (n=1000), determinism check, RTB failsafe timeline, multi-run overlay comparison, fault-tolerance analysis, uncertainty quantification, a self-check audit tab, and an illustrative safety-case outline (not an external compliance assessment). Companion Python research platform: 42 files, 98 tests, 7 experiments, a TLA+ bounded model check of the HMAA authority model (23,748 distinct states, depth 9, no counterexample; 4 of 8 properties exercised, 4 vacuous at these bounds; continuous behavior between decision instants not covered), Dempster-Shafer fusion engine.
Browser implementation of the deterministic authority recovery protocol. Includes the live GREP phase calculator, adversary model analysis, operational replay across six scenarios, parameter space heatmap, and the non-compensatory terminal gate, designed so that a deficit on one criterion cannot be offset by surpluses on others (a design property of the specification).
Real-time sensor attestation engine computing the continuous trust scalar τ ∈ [0,1] consumed by HMAA. Implements TPM-anchored cryptographic chain verification, Monte Carlo trust distribution analysis, FMEA/FTA fault trees, live finite state machine visualization, and DoDAF OV-1 architecture views. All cryptographic operations run in-browser using WebCrypto.
A technical architecture implementing Strategic Latency as an engineered system. Includes a Latency Injection Engine (LIE) middleware, the Dynamic Delay Function D(A, tier, domain), a formal 5-state Circuit Breaker State Machine (NOMINAL, CAUTION, HOLD, FREEZE, LOCKOUT) with hash-logged state transitions (SHA-256 digests; hardware signing is a placeholder in console v5.11, not implemented), Keep-Alive heartbeat protocol with safe-mode link failure defaults, domain-tier risk heatmap, and a simulated physical-interlock reset check.
Swarm trust aggregation with the Byzantine bound as its design basis, extending HMAA to multi-agent environments. Implements trimmed weighted median aggregation (3f+1 Byzantine bound used as a configuration reference), three-layer CUSUM-augmented anomaly detection, graduated escalation with per-level action permissions, DoDD 3000.09 action gate classification, sensitivity analysis, WCET profiling, and a 14-tab PDR briefing package. 37 self-tests run automatically on load. TLA+ specification included; its shipped three-agent configuration exercises no Byzantine agent and marks CriticalSafe, EscalationConsistent and DetectedHandled as expected failures.
A proactive deception prior architecture that adjusts authority downward pre-emptively based on P(adversarial). Implements a Deception Probability Engine computing adversarial likelihood from input distribution anomalies, temporal correlation patterns, cross-sensor consistency, and Bayesian mission history updates. Deception-Adjusted Authority: A_adj = A_hmaa × (1 - λ × P_deception). Includes Phantom Fleet detection module for AI-hallucinated hostile force scenarios.
Strategic monitoring layer that quantifies escalation risk across interconnected autonomous command nodes spanning defense, automotive, maritime, and critical infrastructure domains.
Quantitative cross-domain escalation risk framework for AI-enabled command and control. Computes Decision Compression Ratio (DCR), Authority Chain Integrity (ACI), Cascade Risk Index (CRI), Escalation Probability P(esc), and Human Recovery Window (HRW) across interconnected multi-domain autonomous nodes in real time. AUTHREX tactical command interface with boot sequence, AUTHREX Triforce logo, hex grid background, and CRT scanline overlay. Live ERAM pipeline monitor overlay tracking SATA, HMAA, CARA, MAIVA, FLAME, and ADARA status across all connected nodes. Six scenarios spanning Flash War multi-domain engagement, AV intersection conflict, maritime GPS spoofing, infrastructure cascade, cross-domain defense-to-civilian bleed, and HOTL override under extreme decision compression (DCR=200). 600 Monte Carlo simulation runs with seeded Mulberry32 PRNG (0x4F7A2C1E) for configuration-exact replay within the published console. Real-time 3D topology visualization using Three.js with dynamic node connections, cascade propagation, and Byzantine injection attack testing.
Twelve hardware platforms: two physical research testbeds whose build is in progress (the rover and the UAV) and ten domain reference designs that have not been built. Each published with full BOM, schematics, and interactive simulators.
Controlled experimental platform for UAV flight autonomy under adversarial and degraded conditions. Executes the complete SATA trust fusion, HMAA authority computation, command gating, and CARA recovery pipeline in real-time. Includes GPS spoofing/jamming, RF loss, camera obscuration, and compound attack scenarios, with an experimental MAVLink HIL packet bridge intended for Cube Orange+ hardware integration (not yet tested with hardware; UAV build in progress). Five adversarial scenario families; legacy aggregate run-count claims are withdrawn.
Controlled experimental platform for rover autonomy under adversarial and degraded sensor conditions. Executes the complete 8-stage SATA trust fusion, HMAA authority computation, and CARA recovery pipeline in real-time with configurable fault injection. Seven adversarial scenarios demonstrated in simulation. TLA+ bounded model check of the HMAA authority model (23,748 distinct states, depth 9, no counterexample; 4 of 8 properties exercised, 4 vacuous at these bounds). Rover build in progress.
Complete 9-module defense governance pipeline: SATA → ADARA → IFF → HMAA → MAIVA → FLAME → CARA → BDA → EFFECTOR. Multi-target tracking with Hungarian algorithm WTA optimizer assigning laser, kinetic, and dazzle effectors. 6 threat scenarios with real-time beam suitability (β_beam), 3D collateral clearance, thermal management, and battle damage assessment. Hardware reference design (design package and bill of materials; not built): 72 components; a MIL-STD-810G (superseded by MIL-STD-810H) environmental design target (not tested, certified or assessed).
9-module governance pipeline for autonomous vehicle drive-by-wire control: SATA → ADARA → IFF → HMAA → MAIVA → FLAME → CARA → BDA → EFFECTOR. Three-leg redundant KILOVAC fail-safe relay model. 12 attack scenarios including adversarial ML, V2X spoofing, and compound attacks. twelve attack scenarios demonstrated in simulation. Seeded PRNG for configuration-exact replay within the published console. ISO 26262 ASIL-D target architecture (design target; not tested, certified or assessed).
9-module maritime governance pipeline with four domain-specific extensions: Dempster-Shafer fused trust (hydroacoustic + MAD), recursive AIS phantom vessel deception detection, continuous sea-state authority damping α(H), and acoustic-delay-aware Byzantine consensus. 84-component hardware specification. Split-screen mission map (GIUK Gap). 13 fault injection scenarios. HITL WebSocket bridge (MAVLink/ROS 2).
9-module governance pipeline for ICS/SCADA protection with vintage CRT terminal live mode. 92-component hardware specification. Power grid fault (IEC 61850 GOOSE), water treatment overdose (Modbus TCP), and pipeline leak (Modbus RTU) scenarios. Sensitivity analysis, fault tree analysis, and MITRE ATT&CK for ICS mapping.
Space Authority Protocol simulator (SAP-INT v3): a seven-stage orbital authority pipeline (SATA → ADARA → MAIVA → FLAME → CARA → HMAA → ERAM) with Keplerian-with-J2 propagation, PBFT-style consensus and a SHA-256 hash-chained, ECDSA-signed audit ledger, over six orbital scenarios (LEO conjunction, direct-ascent ASAT threat response, co-orbital approach, multi-stakeholder debris engagement, mega-constellation, Cosmos 1408 retrospective) with fault injection. Companion to the BLADE-SPACE reference design (91 components, $505K BOM, not built); it does not model that design’s 9-stage hardware pipeline.
9-stage AUTHREX pipeline (SENSE → SATA → ADARA → IFF → HMAA → MAIVA → FLAME → ERAM → CARA) arbitrating federal-SLTT authority for Counter-Unmanned Aircraft Systems operations. Four-tier HMAA (T3/T2/T1/T0), Dempster-Shafer consensus across radar, RF, EO/IR, Remote ID, and LIDAR. Six scripted scenarios plus free-play: compliant commercial drone, spoofed Remote ID, motorcade fixed-wing UAS, false-positive bird flock, coordinated swarm probe, and ambiguous track. Deterministic seeded PRNG and SHA-256 audit chain; no aggregate Monte Carlo result is claimed because no run-level result file is published.
Adversarial high-assurance browser emulator for the hardware root of trust that anchors autonomous AI agents, the hardware companion to the AUTHREX-AGENT software shim. Real Web Crypto (ECDSA P-256/P-384, SHA-256, HKDF); PCR measurement chains; non-exportable-key audit signing; four-tier authority (T3 green / T2 amber / T1 red / T0 lock); HKDF per-tool tokens; spawn quorum; multi-modal tamper cascade. Failure-mode scenarios include voltage glitch, electronic-warfare hazard with recovery, split-brain ledger reconciliation, and Byzantine fault isolation. 275 deterministic checks passed in the bundled test report (275/275, seven test batteries), with a software-only-versus-HSM baseline and a P-384 signed golden-trace anchor. Exports a signed evidence bundle; runs a browser self-test.
Discrete-event browser simulator for authority-governed decentralized swarm consensus, parameterised over N=10 (proposed testbed design, not built), N=50 (combined operation), and N=500 (large-swarm scale). Each agent runs a Byzantine-fault-tolerant two-phase consensus gated by SATA peer trust, four-tier HMAA authority, and weighted MAIVA voting before the swarm commits; the protocol is designed to tolerate up to f=(N-1)/3 compromised agents per quorum with a quorum-intersection bound (the Byzantine bound is an assumption of the specification, not a result). Five scripted scenarios from nominal through single-Byzantine, Sybil probe, contested-RF, and denied/degraded. Live seven-stage pipeline, per-node tier array, and a hash-chained distributed audit ledger. Real Web Crypto (ECDSA P-256, SHA-256); safe-halt-by-default under denied RF. Its TLA+ specification states 5 safety and 3 liveness properties; model-checking output has not yet been published.
Deterministic browser simulator for authority-governed cross-boundary IT/OT command adjudication. Each message crossing the IT/OT boundary is parsed, scored against the active authority regime, and adjudicated to propagate, hold for deliberation, or isolate. Four OT authority regimes (NOMINAL, ELEVATED, LOCKDOWN, SAFE-HALT); malformed input fails closed by default. Inject malformed and Byzantine forged-authority traffic, apply clock drift and operator-clearance delay, export CSV metrics, and verify the seed-deterministic SHA-256 hash-chained audit ledger for tampering. Real Web Crypto SHA-256.
Deterministic browser simulator for authority-governed financial-sector AI decision systems. Each transaction is scored through the eight-stage AUTHREX pipeline (VALIDATE through CARA) and routed to one of four authority tiers: autonomous clearance, supervised review, elevated confirmation, or manual hold. Six scenarios including a deepfake authentication attack, an AI-agent coordinated attack, and a low-and-slow ring recovered by the retrospective Swarm Review module. Real Web Crypto SHA-256 over canonical-form serialization, Wilson-interval triage metrics, a Monte Carlo runner, and nine self-tests. Synthetic data only.
These simulations demonstrate the published architectures and research frameworks. HMAA, CARA, SATA, MAIVA, FLAME, ADARA, ERAM, HMAA-UAV, Rover Testbed, BLADE-EDGE, BLADE-AV, BLADE-MARITIME, BLADE-INFRA, BLADE-CUAS, and BLADE-INFRA-OT demonstrate mechanisms described in published specifications and research architectures, with the known deviations noted above. BLADE-SPACE runs the SAP-INT v3 orbital authority simulation, a companion to the Preliminary Design Phase data package (TRL 2-3, Zenodo DOI 10.5281/zenodo.20183269); it does not model that package’s 9-stage hardware pipeline. BLADE-AGENT-HSM runs an adversarial high-assurance emulator whose bundled test report records 275 of 275 deterministic checks passed (Zenodo DOI 10.5281/zenodo.20299821). BLADE-SWARM runs a discrete-event swarm-consensus simulator parameterised over N=10/50/500 with five scenarios (Zenodo DOI 10.5281/zenodo.20351198). BLADE-FINANCE runs a deterministic eight-stage governance simulator over a synthetic transaction stream with six scenarios and a SHA-256 hash-chained audit ledger, on synthetic data only (Zenodo DOI 10.5281/zenodo.20374692). All run entirely client-side with real-time computation.
View Evaluation Protocol ↗, Scenarios, metrics, baselines, and reproducibility
Twenty interactive consoles and demonstrators hosted on the AUTHREX Systems venture site: thirteen application consoles, one per software system, from governor consoles and prototypes to the browser simulation of the AUTHREX-DEFEND product, and seven joint-command and multi-domain demonstrators. Distinct from the nineteen demonstrations hosted on this site. Governance and assurance only; synthetic or invented data; single-author research, not deployed.
Standalone mission simulation of authority lifecycle governance for agentic AI, applied to autonomous cyber defense.
Open Simulation ↗ View Application ↗Cyber-defense console: authorize an autonomous patch by provenance, consistency, quorum, and target criticality.
Open Simulation ↗ View Application ↗Assurance console: walk a candidate through the eight-gate ladder to a signed CONDITIONS MET or HELD IN TEST decision.
Open Simulation ↗ View Application ↗Synthetic T&E console demonstrating how registered actions are evaluated, constrained, recorded, and reset.
Open Simulation ↗ View Application ↗IT/OT boundary console: attest, scope, window, and arm rollback for AI-origin commands before the safety layer.
Open Simulation ↗ View Application ↗Orbital console: onboard authority decisions under light-speed delay, from EXECUTE ONBOARD to SAFE-HOLD.
Open Simulation ↗ View Application ↗Run-time-assurance research console: an independent invariant monitor over a deterministic seven-stage pipeline.
Open Simulation ↗ View Application ↗Engagement-authority research console adjudicating each proposed action across five outcomes, hold-for-human included.
Open Simulation ↗ View Application ↗Safe-state research console governing safe-state entry, hold, deepen, and the re-arm gated exit.
Open Simulation ↗ View Application ↗Team authority 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 and a safe-hold fallback.
Open Simulation ↗ View Application ↗Positive human control prototype: the critical authorization can be held only by two distinct authenticated humans, a machine is never eligible, and the fail-safe state is NO-GO, with a hash-chained record and deterministic replay.
Open Simulation ↗ View Application ↗Agent swarm containment prototype: when one of eight governed AI agents proposes an action outside its grant, a modeled host gate denies it before execution and opens a case, and a new grant requires three distinct witness ballots, five isolation planes, two custodian approvals, a clean dwell and read-only probation, with every decision appended to a hash-chained journal. Synthetic data, no enforcement, TRL 2 to 4 self-assessed.
Open Simulation ↗ View Application ↗The browser simulation of AUTHREX-DEFEND, the installable engineering preview offered to United States Government organizations only. With invented data, it enrolls a pretend program, freezes it, opens a separate synthetic case, advances it through OFF, ELIGIBLE, ARMED, ENGAGING, STOPPED and SEALED, and shows the evidence, with a guided demo and a failure lab. It does not contain the product’s four report-only pages and cannot see your computer. A simulation, not the product.
Open Simulation ↗ View Product ↗Three standalone authority-governance consoles with dedicated application pages. Governance and assurance only; no dynamics, actuation, targeting, or weapons scope.
Standalone run-time-assurance governor console on ASTM F3269 and Simplex. An independent invariant monitor evaluates each proposed authorization before it is applied, with a SHA-256 hash-chained decision ledger.
Open Console ↗ Open Simulation ↗Standalone fail-closed engagement-authority console with five outcomes. A fail-closed invariant monitor adjudicates each proposed action to one of five outcomes: authorize, clamp, downgrade, deny, or hold for human.
Open Console ↗ Open Simulation ↗Standalone fail-safe safe-state authority console with a re-arm gated exit. A fail-safe governor adjudicates every safing command (safe-state entry, hold, deepen, and re-arm gated exit) with a SHA-256 hash-chained decision ledger.
Open Console ↗ Open Simulation ↗Seven demonstrators applying the governance frameworks across joint and multi-domain command scenarios. Governance, consensus, and audit functions in synthetic scenarios; client-side and seeded for reproducibility; no targeting or weapons function.
Six scenarios with 600 Monte Carlo runs, sampled property checks (not model checking), and Merkle provenance, modeling cross-domain escalation cascades.
Open Simulation ↗Full seven-framework pipeline with PBFT consensus and a Merkle audit trail in a theater-level scenario.
Open Simulation ↗Air, maritime, and ground lanes with Byzantine fault isolation under HMAA and CARA governance.
Open Simulation ↗Authority-conflict interdiction demonstrated through a governed common operating picture.
Open Simulation ↗Distributed Web Worker nodes, common operating picture, analytics, and a Merkle audit trail in one architecture.
Open Simulation ↗Actor-model mesh with MAIVA trust aggregation and an emergent interlock across isolated nodes.
Open Simulation ↗Zero-dependency kinematic engine with flight-termination authority under CARA and HMAA governance.
Open Simulation ↗