Skip to main content
Use this page to route review questions to the correct Canon document.

Employee review

Read: Use when the question is about simple reviewer workflow, Quick Review, sliders, or employee onboarding. Also read Employee Onboarding Gate before treating any employee as ready for live evaluator work.

Senior review / adjudication

Read: Use when the question is about final calls, canon candidates, reusable learning, or label meaning.

Exports / schema

Read: Use when the question is about what fields exist and how to interpret exported evidence.

Behavioral Observatory labels

Read: Use when the question is about saved intent, guest affect, response strategy, humanness, notes, saved/unsaved/error state, or the public.eval_behavioral_labels table. Do not use Registry Diagnostics as proof that a Behavioral Observatory label was saved.

Registry Diagnostics

Read: Use when the question is about derived dataset membership suggestions, confidence, source fields, Human Review Queue 2.0 lane suggestions, or Noise / Watch patterns. Do not use Registry Diagnostics as a label workflow.

AI-reviewed platform evidence

Read: Use when the question is about platform lifecycle readiness, Supabase count truth, localStorage compactness, Run History, or Analysis. Do not use these pages as proof of human Lucia-quality approval.