Published May 26, 2026 | Version 0.1.1

RCEA Passport Engine: Role-Conditioned Evidentiary Adequacy Reference Implementation

Authors/Creators

  • 1. AffectLog / Provena

Description

Reference implementation of Role-Conditioned Evidentiary Adequacy (RCEA), a computable framework for assessing whether AI audit evidence is adequate for role-conditioned governance decisions.

The engine generates role-specific evidence passports with deterministic scoring, full traceability, suppression logging, and limitation propagation from a common synthetic evidence substrate. Companion to the manuscript: Role-Conditioned Evidentiary Adequacy for AI Governance: A Computable Framework, Evidence Passport Architecture, and Reference Implementation.

Includes: 7-subscore RCEA engine (Material Relevance, Epistemic Warrant, Normative Alignment, Interpretive Fit, Decision Actionability, Limitation Propagation, Audit Traceability), 4 synthetic worked examples across privacy, robustness, supplier, and fairness domains, 60 passing tests, JSON Schema definitions, CLI scripts, paper artifact reproduction scripts, and 2 Jupyter notebooks.

Warning: This is a research reference implementation using synthetic data only. It is not a legal compliance certification engine.

Files

rcea-passport-engine-0.1.1.zip

Files (98.8 kB)

Name Size
md5:086865045470461fcabfde6d07ba8627
98.8 kB Preview Download

Additional details