Governance evidence engine for ML-based small business credit underwriting: synthetic monitoring, adverse-action reason QA, fair-lending screening, and model-risk oversight.

AGPL-3.0 reference impl of HR Tech audit-stream. Only Suite audit-stream with TWO truly orthogonal candidate-protection invariants: human-hiring-decision (EEOC + Title VII) + NYC LL 144 candidate-notice with 14-day backward window. Runs MomentumHR × HireAssess v2.x end-to-end.

Runnable examples for the CipherExplain Python SDK — encrypted SHAP attributions, ECOA Form C-1 counterfactuals, attestation verification, DP-SHAP, MLP CKKS, tree-attested SHAP, etc.