harn sync

harn sync <subcommand> [options]

Keep a curated subset of .harnery/ live across machines via rclone. Google Drive is the expected remote, but any rclone backend works.

Doesn’t wrap OAuth: rclone config handles Google Drive setup, and harn sync init then picks the configured remote and records it.

What syncs

A deliberately small subset:

• .harnery/identities/: durable persona UUIDs + machine labels
• .harnery/journal/archived/: historical journals
• .harnery/councils/: manifests + invites + per-round contributions

What stays put (machine-local, by design):

• .harnery/active/: live heartbeats
• .harnery/pid-map/: ppid → owner_id lookups
• .harnery/events.ndjson: append-only event stream (large + machine-specific)
• .harnery/.last-intent.*: per-turn intent stamps

Subcommands

harn sync init [--remote <name>] [--prefix <path>]
harn sync status [-v]
harn sync push [--dry-run -v]
harn sync pull [--dry-run -v]
harn sync list [--depth <n>]

Setup

# One-time: configure an rclone remote (interactive OAuth for Google Drive)
rclone config

# Tell harnery which remote to use (records to ~/.config/harnery/sync.json)
harn sync init --remote gdrive --prefix harnery/my-monorepo

# Verify the connection
harn sync status

# First push
harn sync push -v

Day-to-day

# On machine A: capture today's journal archive + council work
harn sync push

# On machine B: pull A's state down
harn sync pull

Defaults

Setting	Default
Remote	(must set)
Prefix	harnery

Env overrides: HARNERY_SYNC_REMOTE, HARNERY_SYNC_PREFIX.