Resolve SCE - handover ecosystem
Everything orbits the working prototype. Ed clicks through it, dev team rebuilds from it, PRDs describe it. Dashed amber borders = gaps to fill. Dashed red border = migration items for dev team handover. Green borders = live and accessible now.
Central artifact - everything radiates from this
Lovable prototype (the working product)
The validated spec. Every screen a seller touches, every decision a broker makes, every credit spent. Ed sees this and clicks through it. Dev team rebuilds from this. PRDs describe what this does. Architecture diagrams map what's inside it. The prototype IS the product - everything else supports it.
Ecosystem setup in progress. Design system repo initialized, Supabase connected, GitHub repos live. Storybook deploying tomorrow. First real screen (E02-S00 platform shell) is the next build step.
↓ built from / validated by
Code (GitHub)
resolve-design-system repo
resolve-seller-connect-engine repo
CLAUDE.md (agent instructions)
Commit history + Layer 2 logs
Backend (Supabase)
Project live
Auth (magic link)
Database (all tables)
Storage (files, audio)
Edge functions (relay, scoring)
Design system (single source of truth)
shadcn/ui components (14 base)
globals.css tokens
Storybook URL (deploying tomorrow)
.lovable/rules/ folder
shadcn/studio Figma kit
Code Connect mappings
Documentation (Google Drive + project knowledge) - 8 PRDs, 102 stories
E1 Legal PRD (7)E2 Intake PRD (11)E3 Screening PRD (7)
E4 Matching PRD (9)E5 Credit PRD (7)E6 Comms PRD (6)
E12 Analytics PRD (12)E14 Emails PRD (5)
Operational system (living)Tech stack registry
Coverage mapCopy guidelines
AI opportunity mapProduction build v3.2
Deal sheet strategyOperational playbook
External services
SendGrid via Twilio (email)
Twilio Voice (masked relay)
DocuSign (eSign)
Stripe (credits + commission)
Mixpanel (analytics)
Anthropic API (TL;DR, scoring)
ABR API (ABN verification)
Retell AI (voice eval)
Migration: Lovable environment to dev team
Lovable holds context that doesn't push to GitHub automatically. When the dev team takes ownership, these need to transfer to their working environment. This is the gap between "prototype works" and "dev team can rebuild from it."
.lovable/rules/ folder (component + pattern specs)
Lovable Knowledge base entries (project context pasted per session)
AI agent configurations (grounding rules, system prompts, what context gets injected)
Lovable project settings (DS connection, GitHub sync config, workspace config)
Prompt patterns that worked (what to paste, what order, what to avoid)
Build session logs (Layer 2 on Trello cards - decisions made during build)
Prototype env vars (Supabase keys, API tokens, feature flags)
shadcn/studio Pro license (transfer or share credentials)
Claude project knowledge files (PRDs, guidelines, architecture context)
GitHub captures code. Supabase captures data. But the AI tooling context - the rules, the grounding, the patterns that make Lovable and Claude Code produce correct output - lives outside both. This needs a dedicated migration doc written after the prototype is proven.
↓ dev team receives all of the above, plus
Dev team receives
Working prototype to click through
Storybook URL (component catalog)
GitHub repos (DS + prototype)
Supabase project (schema + edge functions + data)
Handoff briefs per epic (Layer 3 on Trello)
Architecture diagrams
Test scenarios per story
Tech debt inventory + rebuild guidance
Prototype vs production gap tables
Lovable migration pack (rules, agents, patterns, knowledge)
Layer 3 handoff briefs (written after stories ship)
Lovable migration doc (written after prototype proven)
Last updated 25 March 2026. Update this doc before every meeting with Ed. The prototype URL is the front door - everything else explains what's behind it. The migration section gets its own detailed doc after the prototype is proven and before dev team starts rebuilding.