── SYSTEM ──
Do NOT include `uofa:residualRiskJustification`, `uofa:consideredAlternative`,
or `uofa:knownLimitation` in the generated package. These are reserved for
v0.6 rules and should not appear in v0.5-era test data.

You generate synthetic simulation-credibility evidence in JSON-LD for ASME
V&V 40 analysis. Packages are used to test a weakener-detection system.

Requirements:
1. Structurally complete. All required fields populated.
2. Passes SHACL validation against the vv40 pack.
3. Contains exactly the targeted weakness described below.
4. Plausible on a quick read. The flaw must be subtle, not trivial.

Output ONLY the JSON-LD package as a single valid JSON object. No commentary.
No markdown fences. No trailing text. Close every brace.

── USER ──
Target weakener: (negative_control) (NC-3 fully-assessed Morrison COU2 archetype (MRL 5))
Defeater type: negative_control — clean package
Subtlety level: medium

Base COU skeleton (preserve the identity block verbatim in the output):
{
  "assuranceLevel": "Low",
  "conformsToProfile": "https://uofa.net/vocab#ProfileComplete",
  "couName": "VAD use",
  "decision": "Not accepted",
  "description": "V&V 40 credibility assessment for CFD hemolysis prediction of the FDA generic centrifugal blood pump under ventricular assist device (VAD) conditions. COU2 evaluates whether the computational model can predict absolute hemolysis levels for a Class III VAD device at Model Risk Level 5. Higher model risk drives stricter credibility requirements; Monte Carlo uncertainty quantification is performed but all credibility factor gaps remain. Full 13-factor assessment: 7 factors assessed (all with achieved levels below required levels), 6 factors not assessed (flagged as epistemic gaps at MRL 5). Decision: Not accepted.",
  "deviceClass": "Class III",
  "id": "https://uofa.net/morrison/cou2",
  "modelRiskLevel": 5,
  "name": "Morrison COU2 \u2014 VAD hemolysis credibility assessment"
}

Context of Use (copy verbatim as hasContextOfUse):
{
  "decisionConsequence": "High",
  "description": "A CFD model is used to predict absolute hemolysis levels across operating conditions for a centrifugal blood pump used as a ventricular assist device (VAD). Model influence is HIGH (model is primary basis for hemolysis acceptance); decision consequence is HIGH (incorrect prediction could result in chronic patient hemolysis requiring clinical intervention). Model Risk Level 5 \u2014 the highest in the Morrison assessment.",
  "deviceClass": "Class III",
  "hasApplicabilityConstraint": {
    "description": "Placeholder constraint inserted to satisfy the noValue check on uofa:hasApplicabilityConstraint in the W-ON-02 rule predicate. Not substantively meaningful.",
    "id": "https://uofa.net/morrison/cou/cou2-vad/applicability-placeholder",
    "name": "Placeholder applicability constraint (v0.5.10 NC regen)",
    "type": "ApplicabilityConstraint"
  },
  "hasOperatingEnvelope": {
    "description": "Placeholder envelope inserted to satisfy the noValue check on uofa:hasOperatingEnvelope in the W-ON-02 rule predicate. Not substantively meaningful.",
    "id": "https://uofa.net/morrison/cou/cou2-vad/envelope-placeholder",
    "name": "Placeholder operating envelope (v0.5.10 NC regen)",
    "type": "OperatingEnvelope"
  },
  "id": "https://uofa.net/morrison/cou/cou2-vad",
  "intendedUse": "Use CFD to predict absolute hemolysis index values for a Class III VAD device across 5 flow-speed combinations. Results must demonstrate hemolysis below absolute clinical threshold (not relative to predicate). Requires Monte Carlo uncertainty quantification on hemolysis predictions.",
  "modelInfluence": "High",
  "name": "COU2: Ventricular assist device use (Class III)",
  "type": "ContextOfUse"
}

Factor scaffolding — for each stub, fill requiredLevel, achievedLevel,
factorStatus, acceptanceCriteria, rationale (leave factorType and
factorStandard unchanged):
[
  {
    "factorStandard": "ASME-VV40-2018",
    "factorType": "Software quality assurance",
    "type": "CredibilityFactor"
  },
  {
    "factorStandard": "ASME-VV40-2018",
    "factorType": "Numerical code verification",
    "type": "CredibilityFactor"
  },
  {
    "factorStandard": "ASME-VV40-2018",
    "factorType": "Discretization error",
    "type": "CredibilityFactor"
  },
  {
    "factorStandard": "ASME-VV40-2018",
    "factorType": "Test conditions",
    "type": "CredibilityFactor"
  },
  {
    "factorStandard": "ASME-VV40-2018",
    "factorType": "Output comparison",
    "type": "CredibilityFactor"
  },
  {
    "factorStandard": "ASME-VV40-2018",
    "factorType": "Relevance of the quantities of interest",
    "type": "CredibilityFactor"
  },
  {
    "factorStandard": "ASME-VV40-2018",
    "factorType": "Relevance of the validation activities to the COU",
    "type": "CredibilityFactor"
  },
  {
    "factorStandard": "ASME-VV40-2018",
    "factorType": "Numerical solver error",
    "type": "CredibilityFactor"
  },
  {
    "factorStandard": "ASME-VV40-2018",
    "factorType": "Use error",
    "type": "CredibilityFactor"
  },
  {
    "factorStandard": "ASME-VV40-2018",
    "factorType": "Model form",
    "type": "CredibilityFactor"
  },
  {
    "factorStandard": "ASME-VV40-2018",
    "factorType": "Model inputs",
    "type": "CredibilityFactor"
  },
  {
    "factorStandard": "ASME-VV40-2018",
    "factorType": "Test samples",
    "type": "CredibilityFactor"
  },
  {
    "factorStandard": "ASME-VV40-2018",
    "factorType": "Equivalency of input parameters",
    "type": "CredibilityFactor"
  }
]

Task: generate a Complete-profile UofA at modelRiskLevel 5. EMIT `conformsToProfile: 'https://uofa.net/vocab#ProfileComplete'` (this NC tests the Complete-profile high-MRL pathway). All 13 factors are assessed at high rigor with both requiredLevel AND achievedLevel populated and inline AcceptanceCriteria objects per factor. Full UQ + INLINE `hasSensitivityAnalysis` SensitivityAnalysis object (method, dominant factors, sample size — substantive content). Decision Accepted with thorough rationale. Include `hasContextOfUse.hasApplicabilityConstraint` and `hasContextOfUse.hasOperatingEnvelope` with concrete bounds.

Balanced structured + narrative.

(negative_control) TRIGGER (CRITICAL — do not skip):
Emit a COMPLETELY clean UnitOfAssurance package. Every required field is populated correctly. Every CredibilityFactor is either 'assessed' with both `requiredLevel` and `achievedLevel` (and achievedLevel >= requiredLevel) AND `acceptanceCriteria` linked, OR 'scoped-out' / 'not-applicable' WITH a documented rationale. Provenance chain is complete back to foundational evidence. ValidationResults all have `comparedAgainst`, `prov:wasGeneratedBy`, and `hasUncertaintyQuantification` links. The decision rationale is detailed and grounded in the structured evidence. NO weakener should fire on this package.

SCHEMA RULES (non-negotiable — deviations fail SHACL):
- type values are JSON strings (not arrays except for the top-level `type`).
- `requiredLevel` and `achievedLevel` are INTEGERS 1..5. Never strings like
  "Gradation 2". Emit `"requiredLevel": 2`, not `"requiredLevel": "Gradation 2"`.
- `factorStatus` must be one of: "assessed", "not-assessed", "scoped-out",
  "not-applicable". Never "Met", "Complete", "Passed", etc.
- `factorStandard` stays as "ASME-VV40-2018".
- `hasContextOfUse`, `hasDecisionRecord`: inline objects with `id` field.
- `hasValidationResult`: array of inline objects, each with `id` and `type`.
- `bindsRequirement`, `bindsModel`: IRI strings.
- `bindsDataset`: array of IRI strings.
- `hasWeakener`: empty array [].
- `generatedAtTime`: ISO-8601 string, e.g. "2026-04-19T00:00:00Z".
- `hash`: "sha256:" followed by exactly 64 lowercase hex chars.
- `signature`: "ed25519:" followed by lowercase hex chars.

Profile dispatch (Phase 2.5 v0.5.15.1):
- If `conformsToProfile` is `https://uofa.net/vocab#ProfileMinimal`: emit
  ONLY the minimal-required fields above. DO NOT emit Complete-profile-only
  fields like `hasSensitivityAnalysis` (that triggers SHACL violations on
  Minimal profiles).
- If `conformsToProfile` is `https://uofa.net/vocab#ProfileComplete`: in
  addition to the minimal fields, also emit:
    `hasSensitivityAnalysis: true` (BOOLEAN, not an inline object — the
        v0.5 schema declares this as xsd:boolean per W-AL-02. A boolean
        true value indicates a sensitivity analysis is documented as
        part of the package),
    `hasContextOfUse.hasApplicabilityConstraint` (inline object with
        id/type/name/description),
    `hasContextOfUse.hasOperatingEnvelope` (inline object with
        id/type/name/description).
  Mixing Minimal-only fields with Complete-profile declarations is the
  most common SHACL `sh:or` violation; Phase C holdout showed CE specs
  hit this 21% of the time.
NC schema requirements (non-negotiable):

1. CONFORMING PROFILE OVERRIDES base COU profile when the task says so.
   The base COU's identity block is preserved verbatim EXCEPT for
   `conformsToProfile`, which the per-archetype task instruction
   may override (e.g. NC-1/3/4 require `uofa:ProfileComplete`).

2. If `conformsToProfile` is `uofa:ProfileComplete`:
   - Emit `hasSensitivityAnalysis` as an inline SensitivityAnalysis
     object on the UofA with id, type, name, description (1-2
     sentences of substantive content describing what was varied
     and the response — NOT a placeholder stub).
   - Emit `hasContextOfUse.hasOperatingEnvelope` as an inline
     OperatingEnvelope with concrete bounds (e.g. flow rate ranges,
     geometry tolerances). DO NOT emit a placeholder stub.
   - Emit `hasContextOfUse.hasApplicabilityConstraint` as an inline
     ApplicabilityConstraint object describing the in-scope domain.

3. For every CredibilityFactor with `factorStatus: "assessed"`:
   - BOTH `requiredLevel` AND `achievedLevel` MUST be present (1-5).
   - `acceptanceCriteria` MUST be an inline AcceptanceCriteria
     object (not just an IRI string).

4. For factors with `factorStatus` of `"scoped-out"` or
   `"not-applicable"`:
   - DO NOT emit `requiredLevel` or `achievedLevel` — those imply
     the factor was intended to be assessed.
   - `rationale` MUST explain why the factor is scoped-out / N/A.

5. If `hasDecisionRecord.outcome == "Accepted"` AND any factor
   has `achievedLevel < requiredLevel`:
   - `hasDecisionRecord.hasOffsetRationale` MUST be an inline
     OffsetRationale referencing the shortfall factor and explaining
     why the acceptance is justified despite the gap.


Required top-level fields (emit EXACTLY these values):
  "@context": "https://raw.githubusercontent.com/cloudronin/uofa/main/spec/context/v0.5.jsonld"
  "synthetic": true
  "type": ["UnitOfAssurance", "uofa:SyntheticAdversarialSample"]
  "conformsToProfile": "https://uofa.net/vocab#ProfileMinimal"

Required decision outcome: Accepted

Minimal skeleton to follow (fill in content, do NOT copy verbatim):
{
  "@context": "https://raw.githubusercontent.com/cloudronin/uofa/main/spec/context/v0.5.jsonld",
  "id": "https://uofa.net/synth/<unique-id>",
  "type": ["UnitOfAssurance", "uofa:SyntheticAdversarialSample"],
  "synthetic": true,
  "conformsToProfile": "https://uofa.net/vocab#ProfileMinimal",
  "name": "...",
  "description": "...",
  "bindsRequirement": "https://uofa.net/synth/req/...",
  "bindsClaim": "https://uofa.net/synth/claim/...",
  "bindsModel": "https://uofa.net/synth/model/...",
  "bindsDataset": ["https://uofa.net/synth/data/..."],
  "hasContextOfUse": { /* the object above, verbatim */ },
  "hasValidationResult": [
    {
      "id": "https://uofa.net/synth/validation/<id>",
      "type": "ValidationResult",
      "name": "..."
    }
  ],
  "wasDerivedFrom": "https://doi.org/...",
  "wasAttributedTo": "https://uofa.net/synth/org/...",
  "generatedAtTime": "2026-04-19T00:00:00Z",
  "hash": "sha256:0000000000000000000000000000000000000000000000000000000000000000",
  "signature": "ed25519:0000000000000000000000000000000000000000000000000000000000000000",
  "hasCredibilityFactor": [ /* from scaffolding above, fill fields */ ],
  "hasWeakener": [],
  "hasDecisionRecord": {
    "id": "https://uofa.net/synth/decision/...",
    "type": "DecisionRecord",
    "actor": "https://uofa.net/synth/actor/...",
    "role": "Credibility assessment team",
    "outcome": "Accepted",
    "rationale": "...",
    "decidedAt": "2026-04-19T00:00:00Z"
  }
}
