# Breakout Parity **A testable control architecture for post-deterrence stability** Author: Jonathan Sandhu Public concept-prior date: 2026-07-26 Framework version: 4.0.0-draft-release Literature amendment: 1 Operational claim amendment: 4 Demonstrator amendment: 5 Project codename: Omega Accord Lab Breakout Parity treats nuclear proliferation as a control-system problem. Destroying an inventory or facility does not end proliferation when technical knowledge, motive, restart capacity, weapon-usable material, and political grievance remain. The framework therefore asks whether detection, technical attribution, and effective response can complete before a violator converts breakout into coercive advantage, while compliant states receive a superior package of security, energy reliability, economic access, status, and institutional continuity. The governing control criterion is: ```text T_detect + T_attribute + T_respond < T_breakout_to_coercive_advantage ``` The architecture contains four service-producing operating systems followed by a sovereign constitutional bargain: 1. Decision-time protection, crisis communication, notification, command-system restraint, and material security. 2. A compliance dividend built from assured fuel services, finance, scientific access, conventional protection, and protected rollback. 3. Federated verification, evidence-provenance rules, graded findings, reversible consequences, and automatic restoration after cure. 4. Conversion finance and successor missions for laboratories, production sites, bases, contractors, workers, and host communities. 5. Aggregate military declarations, no-growth ceilings, fissile-production restraint, verified dismantlement, monitored material custody, managed latency, and withdrawal-surviving obligations. The adoption unit is a bounded claim or pilot with a named owner, antecedents, failure condition, receipt, and correction path. Publication does not imply endorsement by any institution named in the framework. ## Current operational finding V10 narrows the architecture to a conditionally sufficient rolling-threat control system inside a qualified operating envelope. It can manage heterogeneous covered threats, decoys, open-world arrivals, sensor disruption, N-1 institutional failure, recurrence, and restoration behavior only after a trustworthy entry baseline has been established. The transition fails when a strategically relevant hidden programme, material path, supplier, sanctuary, or network already exists outside that baseline. It also fails when transformative capability compresses the coercive-advantage clock beyond the maximum qualified sensing, custody, interpretation, response, and reinforcement envelope. The recommended experimental operating gate is: ```text adverse-bound p10 portfolio margin >= 6 months probability of any breach <= 5% restoration on time >= 95% N-1 capacity passed no known uncovered strategic holdout portfolio load inside the certified saturation boundary ``` These are model acceptance thresholds, not empirical estimates of current institutional capability or forecasts of named states. A failed gate stops further reductions and activates hold-and-reinforce procedures. It does not automatically authorize preventive war or permit a claim that operational zero remains stable. ## Current demonstrator status V11 converts the V10 control law into a runnable, role-separated synthetic tabletop at [`demonstrator/`](demonstrator/). It contains seven bounded institutional seats, a persistent threat registry, finite resource queues, N-1 failure, decoys, disputed interpretation, rollback, cure, restoration, authority-overreach rejection, gate evaluation, and canonical receipt export. The retained reference paths behave as required: | Scenario | Required disposition | Retained result | | --- | --- | --- | | Qualified entry with decoy and N-1 loss | Operate after cure, restoration, and recertification | PASS | | Preexisting hidden network holdout | Reveal the uncovered lineage and return to HOLD | PASS | | Ambiguous event under critical-node failure | Preserve dissent and role separation, clear decoys, control the real case, restore, and pass N-1 | PASS | | Portfolio saturation | Hold before optimistic continuation or indiscriminate action | PASS | | Authority-failure path | Reject and record actions attempted from the wrong seats | PASS | V11 demonstrates that the architecture can be represented as an auditable institutional workflow. It does not demonstrate that current institutions possess the required legal authority, data access, throughput, sensing reliability, restoration credibility, or political support. The next falsification step is a role-separated exercise using synthetic or sanitized data and measured interface times. ## What this public record establishes This directory preserves the framework's twelve exact claims, mechanisms, responsible actor classes, first actions, and falsification tests as published on 26 July 2026. It establishes the date and content of Jonathan Sandhu's exact formulation. It does not establish priority over the underlying problems or mechanisms, and it does not establish institutional delivery, acknowledgement, external review, sponsorship, or adoption. The complete qualified V4 archive is identified by this SHA-256 digest: ```text 4a72a71b7c85482fb103c8e7e10e9ca94b0ce1923bd231a457533fd50d155b56 ``` The local V10 qualified-entry archive is identified by: ```text 21452a8c0be8f323ddc5ab189382f84d9ad797d07a1c6cc38ca3191eebbcd284 ``` The binary archives and fixed-layout briefs remain outside this compact host pending dedicated repository and DOI custody. Their absence is recorded rather than concealed. ## Literature status The original capsule followed a targeted institutional-capability and current-policy review, independent systems synthesis, and simulation. It did not follow a field-wide literature review. Literature Status Amendment 1 corrects that boundary without rewriting the historical claim set. The structured first pass covers eighteen literature lanes and 108 anchor sources. Sixty-seven entries were checked through official, publisher, author, or institutional records. Forty-one remain bibliographic seeds pending lawful full-text acquisition and coding. The review produced twenty-four coded contradictions, eighteen model revisions, and six review-derived extension requirements. Every lane remains open for backward and forward citation chaining, contrary-school search, and saturation testing. Strong prior art exists for reconstitution after abolition, post-zero instability, nuclear latency and hedging, fissile-material control, verification of absence and dismantlement, alliance assurance, organizational failure, compliance and enforcement, domestic conversion, and the causes of proliferation and restraint. The candidate contribution is therefore narrower: timed cross-domain integration, persistent rolling-threat control, constrained portfolio capacity, executable restoration, gate-controlled sequencing, and a functional retirement test for mutual assured destruction. The six institutional outreach drafts remain unsent. The public record and runnable demonstrator precede any request for external review. ## Files - `CLAIM_LEDGER.csv`: the twelve stable claims as originally published. - `CONTROL_ARCHITECTURE.md`: the operating model, gate logic, authority triage, and retirement test. - `NOTICE_PROTOCOL_AMENDMENT_1.md`: the evidence ladder separating public existence, attempted transmission, channel acceptance, acknowledgement, review, testing, and formal insertion. - `LITERATURE_STATUS_AMENDMENT_1.md`: the corrected prior-art and novelty boundary. - `PUBLIC_CLAIM_AMENDMENT_4.md`: the V10 qualified-entry, failure-boundary, and operating-gate correction. - `PUBLIC_CLAIM_AMENDMENT_5.md`: the V11 runnable-workflow and next-falsification-step correction. - `V10_DECISION_RECORD.json`: machine-readable V10 decision and evidence boundary. - `V10_RESULT_SUMMARY.csv`: selected V10 entry, validation, red-team, and saturation results. - `V11_DEMONSTRATOR_DECISION_RECORD.json`: machine-readable V11 workflow result and evidence boundary. - `V11_RESULT_SUMMARY.csv`: retained V11 scenario dispositions. - `demonstrator/`: runnable static tabletop, protocol, role cards, acceptance tests, field-data request, and source ledger. - `CLAIM_GENEALOGY_V1.csv`: antecedents, support, contradictions, inherited mechanisms, candidate deltas, revisions, and required tests for all twelve claims. - `CORE_BIBLIOGRAPHY_V1.csv`: public anchor records spanning the literature lanes. - `CONTRADICTION_REGISTER_V1.csv`: serious objections and their resolution experiments. - `MODEL_REVISION_QUEUE_V1.csv`: changes required before the simulator can support operational sufficiency claims. - `EXTENSION_REQUIREMENTS_V1.csv`: literature-derived requirements, identified as `XR-013` through `XR-018` so the original claim ledger remains immutable. - `SOURCE_LEDGER_VERIFIED_2026-07-26.csv`: the original current-institution source ledger. - `RELEASE_RECORD.json`: machine-readable identity, scope, and integrity record. ## Evidence boundary The simulation results are assumption-driven mechanism experiments, not empirical forecasts. Their use is comparative: to expose dependencies, sequencing failures, and conditions under which a regime becomes brittle. The demonstrator scenarios are synthetic. Passing them establishes internal consistency and fail-closed behavior, rather than real institutional performance. The public record requires model ensembles, historical replay, pre-registration, independent replication, and an explicit forecast prohibition before quantitative outputs receive operational weight. ## Control question Can separate real institutions use sanitized data to reproduce the V11 workflow, measure every handoff and queue, preserve role authority and dissent, restore corrected compliance on schedule, and retain a positive adverse-bound margin after a critical-node failure? Documentation is licensed under CC BY 4.0. Code in the complete estate is licensed under MIT unless a file states otherwise.