CROMTEC.AI
Sample GalleryDefenseD3 — Validation & Acceptance Report
🛡Defense · D3 · Veterans Health ProgramAnonymized — Real Engagement

D3 — Validation & Acceptance Report

8 validation scenarios for ATLAS-M readiness intelligence — 6 functional, 2 non-negotiable (VAL-007 Receipt Integrity, VAL-008 Role Restriction). Defense-variant validation includes NASA ERI calculation testing, 3-tier authority chain routing, and DHA compliance gate. Results populate during IMPLEMENT phase.

📄 D3 Deliverable — Validation & Acceptance Report · Engagement ENG-002 · PDIO Phase 3 of 4
⏳ Go-Live Decision: PENDING VALIDATION

Validation Results Summary

ScenarioCategoryResult
VAL-001: Signal Ingestion — 10 SignalsFunctional⏳ Pending
VAL-002: NASA ERI CalculationFunctional⏳ Pending
VAL-003: Composite Readiness ScoreFunctional⏳ Pending
VAL-004: Authority Routing — 3-Tier ChainFunctional⏳ Pending
VAL-005: Receipt Generation — Defense FieldsFunctional⏳ Pending
VAL-006: Escalation TriggersFunctional⏳ Pending
VAL-007: Receipt IntegrityNON-NEGOTIABLE⏳ Pending
VAL-008: Role RestrictionNON-NEGOTIABLE⏳ Pending

Detailed Validation Scenarios

VAL-001

Signal Ingestion — 10 Signals

⏳ Pending

Verify all 10 configured signals transmit data within specified sampling rates: VH-001 fitness per assessment, VH-002 BMI per encounter, VH-003 heart rate continuous (where enrolled), VH-004 PHQ-9 per screen, VH-005 PCL-5 per screen, VH-006 AUDIT-C per screen, VH-007 deployment AQI history (initial + rolling), VH-008 PACT Act flag (record-based), VH-009 current AQI hourly, VH-010 NASA ERI composite daily.

Test Steps
  1. Trigger clinical encounter → verify VH-001, VH-002, VH-004 received within 60s of documentation.
  2. Submit behavioral health screen → verify VH-004/VH-005/VH-006 ingested with correct scale scores.
  3. Verify EPA AirNow polling returns VH-009 within 1 hr window.
  4. Verify NASA Earthdata returns VH-010 ERI composite within 24 hr window.
  5. Verify PACT Act flag ingestion from DEERS record (VH-008) — static and rolling.
  6. Verify normalization: each continuous signal normalized to 0–100 scale.
  7. Verify failover: disconnect biometric monitor → downgrade confidence flag fires within 24 hrs.
Expected Outcome

All 10 signals received within spec. Normalized values within 0–100. Confidence downgrade fires on biometric dropout.

VAL-002

NASA ERI Calculation

⏳ Pending

Verify ERI composite score computed correctly from VH-007 (deployment AQI history), VH-008 (PACT Act flag), and VH-009 (current AQI) per D2 specifications. ERI feeds into GOV-001 readiness score and GOV-002 environmental governance.

Test Steps
  1. Input VH-007 = AQI avg 35 (normal), VH-009 = 42 (normal), no PACT Act flag → verify ERI = low risk.
  2. Input VH-007 = AQI avg 85 (warning), VH-009 = 112 (critical), no flag → verify ERI = moderate.
  3. Input VH-007 = AQI avg 130 (critical), VH-008 = confirmed exposure, VH-009 = 95 → verify ERI = critical.
  4. Verify ERI recalculates within 24 hrs of signal change.
  5. Verify ERI contribution is reflected in composite readiness score for GOV-001.
Expected Outcome

ERI scores match expected values for all 3 conditions. Recalculation within 24 hrs. ERI contributes to readiness composite per D2 weights.

VAL-003

Composite Readiness Score

⏳ Pending

Verify composite readiness score computed from biometric, behavioral, and environmental signal weights per D2. Tier routing (Tier 1/2/3) triggers correctly from score ranges.

Test Steps
  1. Input all signals in normal range → verify composite readiness > 0.75 (Tier 1 routing).
  2. Input PHQ-9 = 12 (moderate) + ERI = 0.45 (warning) → verify composite 0.65–0.74 (Tier 2 routing).
  3. Input PHQ-9 ≥ 15 (critical) &#8594; verify composite < 0.65 (Tier 3 routing + MOR notification).
  4. Verify score recalculates within 60s of signal change.
  5. Verify stale signal flag when any signal > configured staleness threshold.
Expected Outcome

Composite score matches expected values for all 3 conditions. Tier routing fires correctly. Stale signal flagged.

VAL-004

Authority Routing — 3-Tier Chain

⏳ Pending

Verify recommendations route to the correct authority tier per D2 matrix and that Tier 3 generates chain-of-command acknowledgment receipts.

Test Steps
  1. Trigger GOV-001 (composite < 0.75) &#8594; verify routes to AUTH-001 (Primary Care Provider) within 48 hrs.
  2. Trigger GOV-003 (PHQ-9 ≥ 15) &#8594; verify routes to AUTH-004 (Behavioral Health Provider) within 24 hrs.
  3. Trigger GOV-001 (composite < 0.65) &#8594; verify routes to AUTH-002 (Medical Officer of Record) + AUTH-007 notification within 48 hrs.
  4. Let Tier 2 GOV-003 window expire &#8594; confirm auto-escalation to AUTH-005 (Chief of Behavioral Health).
  5. Verify Tier 3 receipt includes time-stamped acknowledgments at each authority level.
Expected Outcome

All 5 rules route to correct tier authority. Auto-escalation fires within SLA. Tier 3 receipt includes full chain-of-command documentation.

VAL-005

Receipt Generation — Defense Fields

⏳ Pending

Verify defense-variant governance receipts contain all required fields including Readiness Tier, ERI Score, LPRM Score, Authority Chain (Tier 1/2/3 acknowledgments), and NASA license references.

Test Steps
  1. Resolve GOV-001 (Tier 2 readiness) &#8594; verify receipt includes ERI Score, LPRM Score, Readiness Tier, and all authority chain acknowledgments.
  2. Verify Decision ID format: GR-YYYYMMDD-VHP-SEQ.
  3. Verify NASA license references (DN-2469, DN-2604, DN-2610, DN-2611) present in ERI-triggered receipts.
  4. Verify confidence within 0.00–1.00.
  5. Verify STATUS = SEALED after final authority signs.
  6. Repeat for GOV-003 (behavioral health escalation) &#8594; verify dual-reviewer receipt fields present.
Expected Outcome

All required fields present including defense-specific fields. Decision ID format correct. NASA licenses present in ERI receipts. Status sealed.

VAL-006

Escalation Triggers

⏳ Pending

Verify confidence threshold escalation: any recommendation with confidence < 0.75 is BLOCKED and escalated. Verify 0.75 threshold (higher than standard 0.70 for military readiness determinations).

Test Steps
  1. Input borderline signals: fitness score 68 (borderline) + PHQ-9 = 6 (mild) &#8594; verify confidence < 0.75.
  2. Verify recommendation BLOCKED and routed to human authority.
  3. Verify receipt shows BLOCKED with confidence value and escalation target.
  4. Input clear high-risk signals: fitness 45 + PHQ-9 = 18 &#8594; verify confidence ≥ 0.75 and recommendation passes normally.
Expected Outcome

Low-confidence blocked and escalated. High-confidence passes normally. Receipt documents block reason.

VAL-007

Receipt Integrity

NON-NEGOTIABLE⏳ Pending

Verify SHA-256 hash computation, receipt chain immutability, and chain break detection. Tests D2 receipt specification. THIS BLOCKS GO-LIVE IF FAILED.

Test Steps
  1. Generate 3 receipts through GOV-001 (Tier 1), GOV-003 (Tier 2), GOV-001 (Tier 3) normal flows.
  2. Recompute SHA-256 of receipt 1 from raw fields in deterministic order &#8594; verify matches stored hash.
  3. Verify receipt 2 chain hash = receipt 1 SHA-256. Verify receipt 3 chain hash = receipt 2 SHA-256.
  4. Attempt modification of Tier 3 receipt authority chain &#8594; verify system rejects.
  5. Verify chain break detection: corrupt receipt 2 hash &#8594; verify escalation fires.
  6. Verify genesis receipt uses GENESIS-VHP chain hash.
Expected Outcome

All SHA-256 hashes match recomputation. Chain links verified. Modification rejected. Chain break detected and escalated.

VAL-008

Role Restriction

NON-NEGOTIABLE⏳ Pending

Verify RBAC prevents unauthorized access to readiness receipts and authority chain data. Specifically tests that Tier 1 providers cannot access Tier 3 readiness determinations and that command-level reports are restricted to authorized authorities.

Test Steps
  1. Log in as AUTH-004 (Behavioral Health Provider) &#8594; verify can view GOV-003 receipts.
  2. As AUTH-004, attempt to view GOV-005 (command readiness report) &#8594; verify ACCESS DENIED.
  3. As AUTH-004, attempt to access Tier 3 readiness determination receipts &#8594; verify ACCESS DENIED.
  4. Verify audit log captures both denied attempts.
  5. Log in as AUTH-002 (Medical Officer of Record) &#8594; verify can view all tiers within their program.
  6. Verify command reporting system access restricted to AUTH-002 and AUTH-007.
Expected Outcome

Role-based access enforced by tier. Cross-rule receipt access denied. Command report access restricted. All denied attempts logged.

LLM-as-a-Judge Test Cases

RuleTest InputExpected Judge Action
GOV-001Fitness score 42 + PHQ-9 = 3 (minimal) + ERI = 0.35PASS Tier 3 — fitness score alone triggers Tier 3 routing
GOV-001Fitness score 88 + all signals normal + ERI = 0.25PASS Tier 1 — routine readiness
GOV-001Fitness score 68 + missing biometric data (30+ days)BLOCK — insufficient data; confidence < 0.75
GOV-002PACT Act flag + deployment AQI avg 145 confirmedPASS — environmental consultation required + PACT documentation triggered
GOV-002Self-reported exposure only (no DEERS record)BLOCK — PACT Act flag requires verified service record, not self-report
GOV-003PHQ-9 = 18 (severe)PASS — Tier 2 escalation to Behavioral Health Provider within 24 hrs
GOV-003PHQ-9 = 4 (minimal) + provider flags AI disagreementPASS — AI-human disagreement documented in receipt per DHA AI Policy requirement
GOV-005Weekly report with unsealed Tier 3 receipt from prior weekBLOCK — all Tier 3 receipts must be sealed before weekly report is generated

Go-Live Authorization Gate

All functional scenarios (VAL-001–006) passed
Yes⏳ Pending
VAL-007 (Receipt Integrity) passed
NON-NEGOTIABLE⏳ Pending
VAL-008 (Role Restriction) passed
NON-NEGOTIABLE⏳ Pending
Judge false negative rate = 0%
Yes⏳ Pending
DHA AI Policy 2023 compliance verified
Yes⏳ Pending
All remediation items resolved
Yes / N/A⏳ Pending

Authorization Requires:

Engagement Lead — sign-off
CISO-CIO — security sign-off
Client Medical Officer of Record — Acceptance sign-off
DHA Representative — AI governance compliance sign-off
← D2 — Governance Design SpecificationD4 — Monthly Governance Report →

Anonymized · Real Engagement · CROMTEC.AI · Patent TPP96862

See what ATLAS would produce for your organization.

Start a conversation to scope your ATLAS engagement.

Start a conversation →← Back to samples