One Kernel. Seven Frameworks.
Six Application Profiles.
AUTHREX uses one Governance Kernel architecture composed of seven governance frameworks. Six of the pages below are application profiles of that architecture, not six independently validated products. AUTHREX-AGENT is the first profile; this page shows the same governed pipeline pointed at documented federal needs. Two lead cases with statutory deadlines, three showing the reach. Three standalone governor consoles, RTA, ENGAGE, and SAFING, ride the same rail, with two hardened prototypes, AUTHREX-TEAM and AUTHREX-REDLINE, beside them. Open any file below to run its live simulation.
Eleven Files. One Pipeline. Pick One.
The product, the five applications, the three standalone governor consoles, and two hardened prototype consoles, each with its gate loadout, its verdict vocabulary, and a live signature scene. Select a file on the rail; open its full page or its standalone simulator.
Evaluate. Assure. Operate.
The set is a lifecycle, not a list. An AI is exercised inside the proposed SANDBOX workflow, evaluated through ASSURE, and, when configured conditions are met, receives an internal release recommendation for consideration by the accountable program authority; any later deployment decision remains external to AUTHREX. It is then modeled under the domain gate that fits its surface.
01 · Evaluate
SANDBOX caps the AI under test at a test tier, rejects action classes configured as prohibited within the simulation and stated mediation assumptions, requests a defined state-reset between runs, and records actions routed through the registered governance interface. The proposed workflow constrains registered actions during synthetic adversarial evaluation; it does not establish complete containment.
02 · Assure
ASSURE walks the evidence through the full gate ladder and either signs an ASSURANCE CONDITIONS MET record to the ledger or returns HELD IN TEST with the failing gate named.
03 · Operate
In a future integrated deployment, an application profile would mediate registered actions at its proposed software enforcement point demonstrated in synthetic simulation (subject to integration completeness and downstream-system integrity): AGENT for agentic runtimes, ICS-GATE at the IT/OT boundary, AGENT-CYBER over autonomous patching, SPACECYBER onboard in orbit.
Same Gates. Same Specification. Different Surface.
Each application profile maps the Governance Kernel architecture to a different domain; the profiles share an authority-state specification, but integration assumptions, evidence, and maturity differ. Model-checking results apply to the stated model and assumptions, not to physical implementations or every domain integration. The HMAA authority state machine underneath is specified in TLA+ and model-checked across the stated finite model; the checker caught a real S5 view-change regression during development.
| File | Role | Gate loadout | Verdict vocabulary | Primary anchor |
|---|---|---|---|---|
| AGENT | Agentic Ai | SATA → ADARA → IFF → HMAA → MAIVA → FLAME → CARA → ERAM | EXECUTE / DELAY / HANDOFF / ABORT | Aligned with CISA + NSA + Five Eyes joint guidance "Careful Adoption of Agentic AI Services" (1 May 2026) |
| ASSURE | Assurance | SATA → ADARA → IFF → HMAA → MAIVA → FLAME → CARA → ERAM | CONDITIONS MET / HELD IN TEST | President Trump’s Cyber Strategy for America (March 2026) |
| ICS-GATE | Operational Technology | SATA → HMAA → FLAME → CARA | EXECUTE / HANDOFF / REFUSED / ABORT | CISA / NSA "Principles for Secure Integration of AI in OT" (3 Dec 2025) |
| AGENT-CYBER | Cyber Defense | SATA → ADARA → MAIVA → HMAA | EXECUTE / HANDOFF / ABORT / HELD | Five Eyes "Careful Adoption of Agentic AI Services" (1 May 2026) |
| SPACECYBER | Orbital | SATA → HMAA → FLAME → CARA | EXECUTE ONBOARD / HANDOFF / SAFE-HOLD | NASA SBIR 2026 BAA subtopic EXPAND.3.S26B |
| SANDBOX | Test & Evaluation | HMAA → FLAME → CARA → ERAM | COMPLETE → ASSURE / CONTAINED | NDAA §1534 (AI sandbox-environments task force, 1 Apr 2026) |
Not Separate Applications, and Labeled as Such.
For honesty about scope: two of the names in the AUTHREX vocabulary are cited variants folded into AGENT-CYBER, and two are features, not standalone applications. They are listed here so the count is never inflated.
AUTHREX-ZTAGENT
Zero Trust for autonomous agents. The same agentic surface as AGENT-CYBER, not a separate application.
AUTHREX-MCPGOV
Model Context Protocol server governance. A citation layer within AGENT-CYBER.
AUTHREX-PQC
Post-quantum-ready signing. A property of how BLADE-AGENT-HSM signs, a hardware feature.
AUTHREX-AISBOM
AI software bill of materials. A signed provenance record the ERAM ledger emits, a ledger feature.
What This Set Is, and Is Not.
These five applications are reference architectures, self-assessed at approximately TRL 2 to 4: specified, simulated, and mapped to publicly deposited hardware reference designs. They are not deployed, not operationally validated, and not adopted or endorsed by any agency. The mapping of each federal need to the AUTHREX pipeline is one researcher’s analytical work, released openly so an independent reviewer can apply the same reasoning and report disagreement. Citations refer to publicly available primary sources. Independent research, offered for technical evaluation, not a contract proposal or a claim of adoption.
CHOOSE YOUR ENTRY POINTSee a signed CONDITIONS MET / HELD decision on a synthetic candidate in under a minute.
OT / UTILITY OWNERGuard the boundaryWatch an AI-origin command attested, windowed, and refused a path around your safety PLC.
CYBER DEFENSE TEAMAuthorize the defenderSame patch, different target: execute on test, handoff on live OT, quarantine when poisoned.
MISSION DESIGNERDecide under delaySet the light-speed delay and see when the vehicle acts onboard versus safe-holds for ground.
T&E ORGANIZATIONContain the evaluationPush an AI-under-test to escape and examine how registered test actions are evaluated, constrained, reset, and recorded within the packaged simulation.
The AUTHREX Application Set · One Kernel, seven frameworks, six application profiles. AUTHREX-AGENT is the agentic-AI product; the other five are application profiles of the same Governance Kernel architecture, composed from the seven AUTHREX governance frameworks.