Validation Plan & Summary Report¶

Document reference: GFTCL-VP-001 · Framework: GAMP 5 Category 5 · EU Annex 11

This document defines how the GaiaFTCL vQbit Quantum VM is validated (the plan) and records the outcome (the summary report). It is the controlling document for the qualification effort.

Part A — Validation Plan¶

Scope¶

The complete bespoke (Category 5) system: vQbit VM, C⁴ algorithm rail, Franklin orchestrator, substrate store, CLI, and the federation surface. The host macOS is Category 1 infrastructure and out of scope except for stated system requirements.

Approach¶

Risk-based validation per GAMP 5 Category & Supplier Assessment and Risk Assessment. The V-model is verified on two independent gates:

Build-time gate — 156 swift test GAMP 5 tests under Tests/GAMP5/ (IQ/OQ/PQ/MQ).
Runtime gate — 18–19 in-app checks (SovereignQualificationRunner) against the live shipping stack inside Franklin.app.

Deliverables¶

URS, FS, SDS, Configuration Spec, RTM, Risk Assessment, IQ/OQ/PQ protocols + reports, Qualification Catalog, and this plan/summary. All are controlled documents in this section.

Acceptance criteria¶

A release is qualified when: all 156 build-time tests pass; all in-app checks return CALORIE; the RTM shows full bidirectional coverage; and every run seals a re-verifiable MQ receipt.

Roles¶

Developer-operator (FortressAI Research Institute) authors and executes; the independent build-time and runtime gates provide testing independence; a third-party assessor can re-verify any receipt without developer access.

Part B — Validation Summary Report¶

Result¶

CALORIE — qualified. The qualification gates pass on the current baseline.

Gate	Count	Result
Build-time `swift test` GAMP 5 suite	156	pass
In-app IQ/OQ/PQ runner	18–19	CALORIE
MQ receipts sealed & retained	120+	append-only

Evidence¶

Each qualification run writes a federation-cosigned receipt to cells/xcode/qualification_receipts/ (MQ-*.json). Each receipt re-computes its witness hash from its sealed canonical witness, so the result is independently reproducible.

Deviations¶

Tracked through Operational Procedures → change control. Any failed check blocks the release until resolved and re-qualified.

Research validation (QC VM + ECDLP + PoW learning)¶

Independent of the in-app qualification runner, the research publication bundle is regenerated from code and sealed JSON:

Artifact	Schema	Role
`ecdlp_research_*`	`shor.ecdlp.research.v2`	Neg-G / structured ECDLP witness
`ecdlp_evaluation_latest.json`	`shor.ecdlp.evaluation.v1`	Three evaluation rows
`qc_vm_validation_latest.json`	`qc_vm.research.v4`	Full QC-001…QC-021 matrix
`qc020_dual_path_latest.json`	`qc020.dual_path.v1`	HIST + LIVE PoW paths
`qc020_improvement_trend_latest.json`	`qc020.improvement_trend.v1`	Time-to-nonce trend

Frozen findings (2026-06-05): allMathLayerOK=true, allLearningClaimsOK=true, qaLayerClosed=true, allRewardClaimsOK=false (L7 not chain_accepted — honest), qc020ImprovementTrend.trendStatus=insufficient_hits until timed hits accumulate.

→ Research-Publication · paper cells/xcode/docs/SHOR_ECDLP_RESEARCH_PAPER.md

cells/xcode/scripts/peer_review_research_dataset.sh

Release statement¶

On a passing baseline the system is released for operational use. Re-qualification is required after any change to a High/Critical configuration item, a substrate schema addition, or a toolchain/OS upgrade.

Cross-references: RTM · Qualification Catalog · IQ · OQ · PQ.

Federation-cosigned

This page's source is sealed in the GaiaFTCL federation manifest — page SHA-256 815b1533298ed581…, manifest witness 3d999b305397d96c…, signed 2026-07-17T15:45:53Z by cell gaiaftcl-mac-cell. Verify with gaiaftcl wiki sign --all and compare wiki-all-signatures.json.