Cross-vault sync
Two vaults exist:
/var/www/kebahagiaan.corone.monster/.wiki/— primary (Julian instance)/var/www/corone.monster/.wiki/— mirror (Ryo instance)
Both must stay byte-identical for shared concepts/entities/synthesis.
What syncs (always mirror keb → corone)
- ✅ All
pages/concepts/*.md(rules + decisions apply globally) - ✅ All
pages/entities/*.md(same projects, same VPS) - ✅ All
pages/synthesis/*.md - ✅
index.md(with workspace name swapped) - ✅
log.md - ✅
check.sh(health check)
What does NOT sync (workspace-local)
- ❌ Per-instance branch-specific notes (rare)
- ❌ Per-instance build logs
- ❌ workspace-tagged sources (paste-in artifacts unique to one instance)
Sync direction
Keb is primary. Mirror keb → corone unless corone has workspace-specific addition (rare).
Sync commands
Standard rsync + sed for index workspace-name swap:
KEB=/var/www/kebahagiaan.corone.monster/.wiki
CORONE=/var/www/corone.monster/.wiki
rsync -a --delete $KEB/pages/ $CORONE/pages/
cp $KEB/log.md $CORONE/log.md
cp $KEB/check.sh $CORONE/check.sh
sed 's|kebahagiaan.corone.monster|corone.monster|g' $KEB/index.md > $CORONE/index.mdWhen to sync
- After every working session that touches concepts/entities/synthesis
- After backfill / lint sessions
- After page-count changes (always update index in both)
Verification
bash $KEB/check.sh && bash $CORONE/check.shBoth must report ✅ count matches + same page count.
Difference in page count = drift, fix immediately.
Common drift causes
- Edit page on one vault, forget to rsync
- New concept added without
cp index.mdswap - Pages dir count vs index.md
Pages: NNmismatch (audit via check.sh)
Skeleton-state failure (avoid)
2026-04-28 → 2026-05-03: skeleton drift caught — keb had 7 pages but index claimed 14, corone had 0. User caught it. Fixed same session by full rebuild + mirror.
Now: every session-end, run check.sh on both vaults. Hard error if mismatch.
Cross-refs
- karpathy-method — vault methodology
- session-start-protocol — boot checklist