Summary

Adds a CLAUDE.md file documenting a repeatable playbook for reorganizing domain modules to match the structure established by dojo/url/. This serves as both developer documentation and AI-assistant guidance for ongoing modularization work.

Motivation

Several core modules (finding, test, engagement, product, product_type) are partially reorganized — UI views, queries, and signals have been extracted to subdirectories, but models (~5k lines), forms (~4k lines), filters (~4k lines), serializers (~3.4k lines), and API viewsets (~3.5k lines) remain in monolithic files. This document codifies the dojo/url/ pattern as the target architecture provides a step-by-step extraction process so contributors can work on modules independently and consistently.

What's included

• Target directory structure matching dojo/url/ with ui/, api/, and shared business logic separation
• 10-phase extraction playbook: models → services → forms → UI filters → UI views/urls → API serializers → API filters → viewsets → API urls
• Service-layer guidance: principles for extracting business logic from views into services.py so both UI and API layers share single implementation
• Per-module audit: specific classes, line ranges, and dependency notes for all 5 target modules
• Cross-module dependency strategy: extraction order and lazy-import patterns to avoid circular imports
• Safeguards: backward-compatible re-exports, admin double-registration prevention, migration verification