project creatortrack public platform main drift 67 commits 20260715T214603

While building CreatorTrack's Social data-layer/Scripts-app and Meta-ingest features across two branches (feat/social-data-layer-and-scripts, feat/social-meta-ingest), Justin's separate "public-platform" workstream (URLs, analytics, funnels, pricing) merged 67 commits to main, including a hardening pass on core create-operation workspace scoping (cf5a4cd, 67e6fcf) and new migrations (workspace-slug registry, scheduling/booking schema).

Both feature branches are build-complete, review-clean, and verified live (typecheck/eslint clean, 497/497 data coverage, live sync re-verified), but as of 2026-07-15 they are not pushed to GitHub and have no PR — deliberately held back so the eventual PR diff isn't tangled with the 67 drifted commits. Rebase analysis (via git merge-tree dry-run simulation) found: Phase 2 rebases onto Phase 1 with zero conflicts; Phase 2 has no impact from the main drift (writes bypass the hardened endpoints entirely); Phase 1 has 6 conflict files against main (1 mechanical, 4 needs-care, docs/APP-PLAYBOOK.md needs human judgment), and Phase 1's create-calls already satisfy main's new workspace contract so no code changes are needed.

Why: Justin flagged the drift himself mid-session and Claude's stated recommendation was not to rebase onto an actively-moving main — wait until the public-platform work reaches a settling point.

How to apply: before resuming this ship, re-check whether main has kept moving; if the public-platform workstream has settled, the rebase-then-push-then-PR sequence is unblocked and pre-analyzed (see conflict list above). Ship-time checklist also includes provisioning SOCIAL_SYNC_CRON_SECRET, enabling the daily-sync timer, restoring transcripts+structures into prod, and Justin reconnecting YouTube on ws 85.

[auto-memory session 60d132de-4659-485c-8044-427af0ccce7f, confidence 0.75, mode staged]