journal-sync

name	journal-sync
description	Pull recent Notion Journal entries into memory/journals/YYYY/month_YYYY-MM_text.txt.
tools	["Bash"]

On load, immediately run:

bash .claude/skills/journal-sync/sync.sh

Do not ask for confirmation — just execute it.

Bootstrap (run once)

If memory/journals/index.md doesn't exist or lacks ## Sync Tracking, create it with this structure so the sync script can track state:

python3 - <<'PY'
from pathlib import Path
p = Path('memory/journals/index.md')
if not p.exists() or not _has_tracking(p.read_text()):
    config = '\n## Configuration\n\n- **Data Source ID**: `cc42b6a6-bb69-481f-bab9-26b22e4b56bc`'
    s = '# Journal Archive Index\n\nIndex and sync state for the journal archive. Monthly entries live under `YYYY/` subdirectories; yearly summaries at root level.' + config + '\n\n## Sync Tracking\n\n> Update this section after each journal-sync. Latest journal date is the persistent high-water mark; lookback catches recently missed entries.\n\n- **Last sync**: never (first run)\n- **Entries fetched**: 0\n- **Entries added**: 0\n- **Lookback days**: 7\n- **Latest journal date**: null (created_time high-water mark)\n\n## Notes\n\n- Monthly raw files live in `YYYY/month_YYYY-MM_text.txt` subdirectories; synced automatically by this skill.\n- Use yearly summaries first for overview, then inspect monthly extracts when exact chronology or wording matters.'
    p.write_text(s)
else:
    s = p.read_text()
    insert = '\n## Sync Tracking\n\n> Update this section after each journal-sync. Latest journal date is the persistent high-water mark; lookback catches recently missed entries.\n\n- **Last sync**: never (first run)\n- **Entries fetched**: 0\n- **Entries added**: 0\n- **Lookback days**: 7\n- **Latest journal date**: null (created_time high-water mark)\n'
    # Add Configuration section if missing.
    if 'Data Source ID' not in s:
        config = '\n## Configuration\n\n- **Data Source ID**: `cc42b6a6-bb69-481f-bab9-26b22e4b56bc`'
        s = s.replace('\n## Sync Tracking', config + '\n## Sync Tracking', 1) if '\n## Sync Tracking' in s else s + config
    p.write_text(s.replace('\n## Notes', insert + '\n## Notes', 1) if '\n## Notes' in s else s + insert)

def _has_tracking(text):
    return '## Sync Tracking' in text
PY

This only runs once. The sync script will then update the tracking state after each run.

Rules

• Incremental state is Latest journal date in memory/journals/index.md.
• Query a small lookback window before that date, default 7 days, so missed recent entries are recovered.
• Deduplicate by normalized first non-empty markdown line before writing, so minor punctuation/whitespace differences do not duplicate entries.
• Do not store or use Notion pagination cursors for routine sync.
• If more than 100 entries match, stop and do a manual backfill.

Notes

• Default lookback means normal daily runs re-fetch a few recent entries, but dedup skips them.
• For a larger catch-up, run with JOURNAL_SYNC_LOOKBACK_DAYS=30 or temporarily lower Latest journal date, then sync again.
• After changes to memory/, append to memory/log.md then trigger memory-backup