Refua Regulatory

Organize research evidence for regulatory review. Link decisions to the plans, tools, models, datasets, and results that support them.

View on GitHub Back to Ecosystem
E2E Lineage
SHA-256 Checksums
FDA Checklist Templates
JSON Structured Output

What Is Refua Regulatory?

Refua Regulatory turns campaign decisions and results into verifiable evidence bundles with end-to-end lineage across plans, tool calls, models, datasets, and generated artifacts.

Every decision in a Refua campaign is extracted, given a deterministic ID, and linked to its model provenance and data manifests. The result is a complete audit trail that supports internal regulatory readiness reviews before agency submission.

What It Provides

  • Campaign decision extraction with deterministic IDs
  • Model and data provenance capture per decision
  • A graph linking decisions to their supporting evidence
  • Evidence bundle packaging with SHA-256 checksums
  • Integrity verification for audit handoff
  • Regulatory checklist evaluation with strict and manual-review gates

Evidence Bundle Contents

📄

Manifest

A structured manifest.json with bundle ID, creation timestamp, source campaign, and included artifact references.

📋

Decision Records

Structured decisions.jsonl with deterministic decision IDs, rationale, model/data provenance, and execution context.

🔗

Lineage Graph

A lineage.json connecting plans to tool calls to artifacts, enabling full traceability of every campaign output.

🔒

Integrity Checksums

SHA-256 checksums for every artifact in the bundle, enabling tamper-detection and verification for audit handoff.

✅

Regulatory Checklists

Automated checklist evaluation against drug discovery comprehensive and FDA/CDER AI-ML templates with strict and manual-review gates.

📦

Artifact Archive

Campaign run files, data manifests, and extra artifacts organized in a standardized directory structure for handoff.

Get Started

Installation

pip install refua-regulatory

Build an Evidence Bundle

refua-regulatory build \
  --campaign-run artifacts/kras_campaign_run.json \
  --output-dir artifacts/evidence/kras_run_001

Verify Bundle Integrity

refua-regulatory verify --bundle-dir artifacts/evidence/kras_run_001

Run a Regulatory Checklist

refua-regulatory checklist \
  --bundle-dir artifacts/evidence/kras_run_001 \
  --template fda_cder_ai_ml \
  --strict

Keep the Evidence Together

Package the research record, check file integrity, and give reviewers a clear path from a decision to its supporting evidence.

View on GitHub Back to Ecosystem