MP06 — Atom Browser
L01 (57)
L02 (42)
L03 (103)
L04 (36)
L05 (103)
L06 (56)
L07 (44)
L08 (40)
MP01 (61)
MP02 (32)
MP03 (65)
MP04 (48)
MP05 (68)
MP06 (45)
45 Atoms
| ID | Name | Family | Definition | File |
|---|---|---|---|---|
MP06-A-AUD-001 |
Audit Coverage Gauge | AUD | What percentage of system components have been audited? | MP06-A-AUD-COVERAGE.json |
MP06-A-AUD-002 |
Audit Independence Gauge | AUD | How independent is the audit process from audited systems? | MP06-A-AUD-INDEPENDENCE.json |
MP06-A-AUD-003 |
Audit Recency Gauge | AUD | How recent are audit findings relative to system changes? | MP06-A-AUD-RECENCY.json |
MP06-A-CMP-001 |
Compliance Rate Gauge | CMP | What percentage of requirements are in compliance? | MP06-A-CMP-RATE.json |
MP06-A-CMP-002 |
Compliance Gap Severity Gauge | CMP | How severe are identified compliance gaps? | MP06-A-CMP-SEVERITY.json |
MP06-A-TIP-002 |
Tip Response Time Gauge | TIP | How quickly are tips acknowledged and investigated? | MP06-A-TIP-RESPONSETIME.json |
MP06-A-TIP-001 |
Tip Verification Gauge | TIP | What percentage of whistleblower tips have been verified? | MP06-A-TIP-VERIFICATION.json |
MP06-AUD-001 |
Audit Criteria Definition | AUD | Defines and maintains audit criteria derived from Charter commitments and ADRs. | MP06-AUD-AUDITCRITERIA.json |
MP06-AUD-003 |
Audit Trail Completeness | AUD | Validates that complete audit trails exist for all framework decisions. | MP06-AUD-AUDITTRAILCOM.json |
MP06-AUD-002 |
Independence Verification | AUD | Ensures audit independence from audited layers to prevent conflicts of interest. | MP06-AUD-INDEPENDENCEVER.json |
MP06-CMP-002 |
ADR Compliance Verification | CMP | Verifies that all implementations comply with approved Architecture Decision Records. | MP06-CMP-ADRCOMPLIANCE.json |
MP06-CMP-001 |
Charter Compliance Check | CMP | Validates compliance with declared Charter commitments across all framework outputs. | MP06-CMP-CHARTERCOMPLI.json |
MP06-CMP-003 |
Governance Process Compliance | CMP | Validates that all governance processes follow declared procedures. | MP06-CMP-GOVERNANCEPROC.json |
MP06-DEC-001 |
Efficacy Pillar Gate | DEC | First of four decision pillars. Requires evidence of meaningful educational impact meeting pre-registered Smallest Effect Sizes of Interest (SESOI) on both primary learning outcomes and humanistic KPI | MP06-DEC-EFFICACY.json |
MP06-DEC-002 |
Fairness Pillar Gate | DEC | Second of four decision pillars. Requires that the Equity Suite standards are met: Algorithmic Fairness Score within bounds, no sustained parity breaches for any context stratum. Linked to continuous | MP06-DEC-FAIRNESS.json |
MP06-DEC-004 |
Governance Pillar Gate | DEC | Fourth of four decision pillars. Ensures the change was made accountably and transparently: auditable chain-of-command, reproducible data traces, clear contestability path. Links to traceability and p | MP06-DEC-GOVERNANCE.json |
MP06-DEC-005 |
Non-Compensatory Synthesis | DEC | Synthesizes four pillar outcomes into final verdict using non-compensatory logic: a gain in one pillar cannot compensate for failure in another. Reflects lexicographic ordering of MP3 North Star where | MP06-DEC-SYNTHESIS.json |
MP06-DEC-003 |
Validity Pillar Gate | DEC | Third of four decision pillars. Enforces Validity Dossier standards: assessments must pass measurement invariance tests, DIF checks, and meet reliability thresholds. Without valid measurement, no cred | MP06-DEC-VALIDITY.json |
MP06-EQT-003 |
Error Parity Gauge | EQT | Measures whether the system's predictive models distribute errors equitably across subgroups. Comprises two sub-metrics: True Positive Rate (TPR) Parity (Equality of Opportunity) ensuring correct iden | MP06-EQT-ERRORPARITY.json |
MP06-EQT-002 |
Exposure Parity Gauge | EQT | Measures whether students across different subgroups have comparable exposure to essential prerequisite content, controlling for the dynamic sequencing of their learning paths. Addresses the risk that | MP06-EQT-EXPOSURE.json |
MP06-EQT-004 |
Latency Parity Gauge | EQT | Measures operational equity by ensuring that the time taken to receive support (automated or human) is comparable across all subgroups. Guards against scenarios where students in disadvantaged context | MP06-EQT-LATENCY.json |
MP06-EQT-001 |
Opportunity Parity Gauge | EQT | Measures whether the probability of receiving a beneficial intervention is equal across different subgroups, conditional on their underlying need. Ensures the system does not systematically overlook o | MP06-EQT-OPPORTUNITY.json |
MP06-EQT-007 |
Context Stratification Controller | EQT | Mandates that all fairness metrics be stratified by context strata defined by Layer 3 (Ecological Context Engine). Prevents Simpson's Paradox where an intervention appears fair in aggregate but harms | MP06-EQT-STRATIFICATION.json |
MP06-EQT-005 |
Parity Threshold Controller | EQT | Defines and manages the configurable parity gap thresholds that trigger automated breach or near-breach actions across all Equity Suite metrics. Analogous to the four-fifths rule in employment law, pr | MP06-EQT-THRESHOLD.json |
MP06-EQT-006 |
Fairness Trade-off Policy | EQT | Encodes the policy for handling mathematically impossible fairness trade-offs. Acknowledges the impossibility theorem (cannot satisfy all fairness metrics simultaneously) and specifies 'equity_first_w | MP06-EQT-TRADEOFF.json |
MP06-GOV-006 |
CARE Principles Alignment | GOV | Ensures alignment with CARE Principles for Indigenous Data Governance: Collective Benefit, Authority to Control, Responsibility, and Ethics. Community oversight functions must respect Indigenous world | MP06-GOV-CARE.json |
MP06-GOV-004 |
Community Oversight Board | GOV | Ensures collective governance through a multi-stakeholder oversight body (similar to IRB/REC) with genuine authority. Representatives from the community actively review system performance and ethical | MP06-GOV-COMMUNITY.json |
MP06-GOV-003 |
Contestability Path | GOV | Ensures a formal mechanism exists for students, teachers, or stakeholders to question and appeal algorithmic decisions. Implements 'voice' from procedural justice theory. Human review must be substant | MP06-GOV-CONTEST.json |
MP06-GOV-001 |
Event Sourcing Integrity | GOV | Verifies that the audit log is implemented as an append-only, immutable event store where every state change is captured as a sequenced event. The current system state is derived by replaying events, | MP06-GOV-EVENTSOURCE.json |
MP06-GOV-005 |
POPIA Compliance Validator | GOV | Validates compliance with South Africa's Protection of Personal Information Act (POPIA). Ensures data subject rights (access, correction, deletion) are implemented. Contestability and appeal pathways | MP06-GOV-POPIA.json |
MP06-GOV-002 |
Traceability Graph Index | GOV | Maintains a rich, interconnected knowledge graph linking WHY (ADRs from MP2), WHAT (releases from MP5), and IMPACT (Accountability Reports from MP6). Enables traversal of entire decision history for r | MP06-GOV-TRACEGRAPH.json |
MP06-MAT-001 |
M0 Foundational Maturity | MAT | Initial maturity stage focusing on establishing core logic and procedures. Audits conducted manually with spreadsheets for parity checks. Quarterly reports. Critical for developing human expertise and | MP06-MAT-M0.json |
MP06-MAT-002 |
M1 Advanced Maturity | MAT | Advanced maturity stage with continuous, automated assurance. Equity Suite and Validity Dossier checks encoded as CI/CD pipeline gates. Orchestrator-led playbooks. Automatic public artefact generation | MP06-MAT-M1.json |
MP06-MAT-003 |
M2 Frontier Maturity - Digital Twin | MAT | Frontier maturity with Assurance Digital Twin capability. Predictive governance through simulation of policy impacts before deployment. Agent-Based Models project outcomes. Counterfactual accountabili | MP06-MAT-M2.json |
MP06-TIP-002 |
TIPS Interpretability Check | TIP | Assesses interpretability of framework outputs for intended audiences. | MP06-TIP-TIPSINTERPRET.json |
MP06-TIP-003 |
TIPS Proactive Disclosure | TIP | Enforces proactive disclosure of limitations, biases, and uncertainties per TIPS requirements. | MP06-TIP-TIPSPROACTIVE.json |
MP06-TIP-004 |
TIPS Stakeholder Communication | TIP | Ensures stakeholder communication meets TIPS standards for clarity and accessibility. | MP06-TIP-TIPSSTAKEHOLDE.json |
MP06-TIP-001 |
TIPS Transparency Assessment | TIP | Assesses framework transparency per TIPS framework requirements. | MP06-TIP-TIPSTRANSPARE.json |
MP06-VAL-007 |
ADR Validity Link | VAL | Ensures every Validity Dossier is perpetually linked to its originating Architecture Decision Record (ADR) from MP2. When an assessment is created or modified, a corresponding validity analysis must b | MP06-VAL-ADRLINK.json |
MP06-VAL-002 |
Configural Invariance Check | VAL | Tests whether the basic factor structure (which items load onto which latent construct) is equivalent across different demographic groups. This is the first level of Multi-Group Confirmatory Factor An | MP06-VAL-CONFIGURAL.json |
MP06-VAL-003 |
Metric Invariance Check | VAL | Tests whether the factor loadings (strength of relationship between items and latent constructs) are equal across demographic groups. This is the second level of MGCFA (weak invariance) and establishe | MP06-VAL-METRIC.json |
MP06-VAL-006 |
Non-Uniform DIF Detector | VAL | Detects Differential Item Functioning (DIF) where an item's bias varies depending on ability level (interaction effect). Uses logistic regression (significance of group × ability interaction) and IRT | MP06-VAL-NONUNIFORMDIF.json |
MP06-VAL-001 |
Reliability Gauge | VAL | Measures the internal consistency and stability of assessment instruments using Cronbach's alpha and McDonald's omega coefficients. An unreliable assessment produces noisy, untrustworthy results. Sets | MP06-VAL-RELIABILITY.json |
MP06-VAL-004 |
Scalar Invariance Check | VAL | Tests whether item intercepts (baseline probability of endorsing an item) are equal across demographic groups. This is the third level of MGCFA (strong invariance) and is the MINIMUM requirement for m | MP06-VAL-SCALAR.json |
MP06-VAL-005 |
Uniform DIF Detector | VAL | Detects Differential Item Functioning (DIF) where an item is consistently easier or harder for one group across all ability levels. Uses logistic regression (significance of group membership predictor | MP06-VAL-UNIFORMDIF.json |