Migrate Second Brain data: luke ? lisa
Purpose: After routing cmo-gpt / MCP to user lisa, existing rows in Supabase may still be keyed luke. This runbook covers Postgres (this repo) and calls out Mem0 Cloud separately.
Postgres (Supabase)
Migration file: knowledge/sales/second-brain/backend/supabase/migrations/017_migrate_luke_to_lisa.sql
What it updates
user_id:patterns,experiences,brain_health,growth_log,review_history,confidence_history,memory_content,examples,knowledge_repo,projects,vault_transcriptsknowledge_repo.author:luke?lisawhere applicable
How to apply
- Prefer
supabase db pushor your usual migration pipeline against the target project (staging then prod). - Alternatively run the SQL in the Supabase SQL editor (single transaction recommended).
Verify
SELECT 'patterns', COUNT(*) FROM patterns WHERE user_id = 'luke'
UNION ALL SELECT 'knowledge_repo', COUNT(*) FROM knowledge_repo WHERE user_id = 'luke' OR author = 'luke';
-- Expect zero rows in prod after migration.Mem0 Cloud
This repository migration does not rename Mem0 external memory. If production Mem0 still stores scopes under luke, choose one:
- Leave legacy scope and rely on new ingest under
lisa, or - Merge / re-tag via Mem0 dashboard or support, or
- Re-ingest critical examples under
lisausing your ingest pipeline.
Document whatever you choose so operators are not surprised by duplicate or stale memories.
Related
knowledge/sales/content/cc-content-system/cmo-gpt/LISA_CMO_GPT.mddocs/plans/2026-05-06-marketing-ops-fractional-cmo-ce-plan.md