Skip to content

Latest commit

 

History

History

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
 
 
 
 
 
 
 
 
 
 

README.md

examples/ — runnable fixtures for Regula

Three minimal projects, one per EU AI Act risk tier, so new users can see what Regula flags without needing their own codebase. Each directory has its own README with the exact output of regula check verified against the current release.

Directory EU AI Act tier Expected finding (with --scope all)
cv-screening-app/ High-risk (Annex III, 4) HIGH-RISK verdict — employment pattern
customer-chatbot/ Limited-risk (Article 50) LIMITED-RISK verdict — chatbot transparency
code-completion-tool/ Minimal-risk clean scan, 0 findings

Quick start

pipx install regula-ai   # 1.7.5 or newer
git clone https://github.com/kuzivaai/getregula.git
cd getregula

regula check examples/cv-screening-app --scope all    # high-risk verdict
regula check examples/customer-chatbot --scope all    # limited-risk verdict
regula check examples/code-completion-tool            # minimal-risk → clean

--scope all is needed because files under examples/ carry example provenance, and the default --scope production excludes their non-minimal findings — a deliberate precision feature (your real scans stay quiet about vendored demo code). The cv-screening project also declares system.domain: employment in its regula-policy.yaml; since v1.7.5 that declaration is what activates the domain-gated employment patterns (verified 16 Jul 2026 against the released 1.7.5).

Try Regula end-to-end in 10 minutes

cv-screening-app/README.md is a complete evaluation journey: every Regula command (check, plan, gap, conform --zip, verify, handoff, bias) against one realistic high-risk fixture, with expected output for each step. End result is the full 26-file Annex IV evidence pack + a portable .regula.zip bundle + a verification report.

Use it when evaluating Regula before running against your own codebase.

Why three fixtures

The EU AI Act sorts AI systems into four risk tiers (prohibited, high-risk, limited-risk, minimal-risk). Each tier has different obligations. Seeing Regula's behaviour across tiers is the fastest way to understand what it will flag on your own code.

regula-rules.yaml in this directory is a separate example showing how to define custom organisation-specific patterns.