ClawHub

Time Clawshine — OpenClaw Time Machine

Data & APIs

Restic-powered encrypted time machine for OpenClaw: hourly incremental snapshots, fast restore by time/snapshot/file, local-only privacy defaults, integrity checks, retention, and optional alerts. Setup can run with sudo, install dependencies, create systemd/cron persistence, restore over current files, and prune or purge recovery data.

Install

openclaw skills install quick-backup-restore

⏱🦞 Time Clawshine — OpenClaw Time Machine

Your agent just nuked its own memory. Now what?

You spent weeks training your OpenClaw agent — building memory, refining context, tuning personality. Then one bad session wipes it. Gone. And your last "real" backup? Yesterday. Maybe last week.

Time Clawshine gives OpenClaw a local encrypted time machine. Every hour, restic takes an incremental snapshot of your agent's brain — memory, sessions, config, everything. Only changed bytes are stored, so backups stay fast and compact. When things break (and they will), you roll back by time, snapshot, or file to exactly the moment before it happened. Not yesterday. Not "the last backup." The exact hour.

Security note: this is a privileged backup/restore tool, not a narrow read-only helper. Setup can install packages and a scheduler with sudo; restore can overwrite current files; retention/prune can remove old recovery points; and optional external integrations can send minimal operational metadata only after explicit opt-in.

One command to install. Zero maintenance. Just works.

sudo bash {baseDir}/bin/setup.sh

Why this exists

ProblemWithout Time ClawshineWith Time Clawshine
Agent overwrites MEMORY.mdHope you saved a copyrestore.sh "2h ago"
Bad session corrupts contextRebuild from scratchRoll back one snapshot
"What changed?"No idearestic diff between any two snapshots
Disk fills upBackup keeps growingDedup — only deltas stored
Something failsYou find out next weekLocal log, optional minimal Telegram ping

What's under the hood

  • Restic — battle-tested backup engine, AES-256 encryption, incremental deduplication
  • 72 snapshots / 3 days of history at hourly resolution (configurable)
  • Disk guard — aborts before filling your disk, optionally alerts via Telegram
  • Integrity checks — automatic restic check every 24 backups
  • Daily digest — Telegram summary with snapshot count, repo size, disk free
  • Healthcheck ping — opt-in healthchecks.io / hc-style endpoint pings (/start, /, /fail) so you also know when the backup stops running, not just when it fails
  • Update awareness — optional ClawHub check, disabled by default, never auto-updates
  • Status dashboard — bin/status.sh for a full health check at a glance
  • Repository cleanup — bin/prune.sh to manually reclaim disk space
  • Self-test — bin/test.sh validates backup→restore→verify roundtrip
  • Guided setup — agent reads SETUP_GUIDE.md and walks the user through every option
  • Dry-run mode — backup.sh --dry-run to validate without writing
  • Uninstall — bin/uninstall.sh for clean removal (preserves data by default)
  • Local-only by default — no Telegram, healthcheck, or update-check egress unless privacy.local_only is explicitly disabled

Technical reference

Repository: configured in {baseDir}/config.yaml Log: configured in config.yaml under logging.file (rotated weekly via logrotate) Password file: configured in config.yaml under repository.password_file (chmod 600 — back this up separately)

When the user asks to set up or install Time Clawshine

First, read {baseDir}/SETUP_GUIDE.md and walk the user through each step interactively. The guide covers Telegram, frequency, retention, extra paths, disk safety, and repo location. Configure config.yaml based on their answers before running setup.

If installing from ClawHub in OpenClaw, use the slug-only form:

openclaw skills install quick-backup-restore

Some OpenClaw CLI versions reject owner-prefixed slugs such as marzliak/quick-backup-restore.

If the user wants a quick install without customization:

  1. Check if already set up: 
    sudo bash {baseDir}/bin/status.sh
  2. Run setup: Preview first if the user wants to see dependencies, system files, and scheduler changes without modifying the machine: 
    bash {baseDir}/bin/setup.sh --dry-run

    sudo bash {baseDir}/bin/setup.sh
    For repo-only setup (no apt-get, no cron, no /usr/local/bin changes): 
    sudo bash {baseDir}/bin/setup.sh --no-system-install
    For CI/automated setup (skip confirmation prompts): 
    sudo bash {baseDir}/bin/setup.sh --assume-yes
  3. Confirm setup succeeded: 
    sudo bash {baseDir}/bin/status.sh

When the user asks to run a manual backup

sudo bash {baseDir}/bin/backup.sh

Then confirm with:

sudo bash {baseDir}/bin/status.sh

When the user asks to check backup status or history

Run the status dashboard:

sudo bash {baseDir}/bin/status.sh

Or show the last 20 log lines:

sudo tail -20 "$(yq e '.logging.file' {baseDir}/config.yaml)"

List all snapshots (most recent first):

sudo bash {baseDir}/bin/restore.sh --help
# Or directly:
REPO=$(yq e '.repository.path' {baseDir}/config.yaml)
PASS=$(yq e '.repository.password_file' {baseDir}/config.yaml)
restic -r "$REPO" --password-file "$PASS" snapshots

Show what changed between the two most recent snapshots:

REPO=$(yq e '.repository.path' {baseDir}/config.yaml)
PASS=$(yq e '.repository.password_file' {baseDir}/config.yaml)
SNAPS=$(restic -r "$REPO" --password-file "$PASS" snapshots --json | jq -r '.[-2:][].id')
restic -r "$REPO" --password-file "$PASS" diff $SNAPS

When the user asks to restore or roll back

Interactive restore (recommended — always dry-runs first):

sudo bash {baseDir}/bin/restore.sh

Restore by time (e.g. "roll back 2 hours"):

sudo bash {baseDir}/bin/restore.sh "2h ago" --target /tmp/tc-restore
sudo bash {baseDir}/bin/restore.sh yesterday --target /tmp/tc-restore

Restore a specific file from the latest snapshot:

sudo bash {baseDir}/bin/restore.sh latest --file /root/.openclaw/workspace/MEMORY.md --target /tmp/tc-restore
# Preview the result, then move manually:
# cp /tmp/tc-restore/root/.openclaw/workspace/MEMORY.md /root/.openclaw/workspace/MEMORY.md

Restore a specific snapshot by ID:

sudo bash {baseDir}/bin/restore.sh <snapshot_id>

Always confirm with the user before executing a full restore to /. For a restore targeting /, the script requires the exact confirmation phrase RESTORE TO /.

When the user asks to check repo integrity

REPO=$(yq e '.repository.path' {baseDir}/config.yaml)
PASS=$(yq e '.repository.password_file' {baseDir}/config.yaml)
restic -r "$REPO" --password-file "$PASS" check

When the user asks to change configuration

Edit {baseDir}/config.yaml with the requested changes (schedule, retention, paths, Telegram credentials), then re-run setup to apply:

sudo bash {baseDir}/bin/setup.sh

When the user asks to customize backup paths

Run the local path analyzer (100% offline — no API calls, no data leaves the machine):

sudo bash {baseDir}/bin/customize.sh

This scans the system for:

  • Extra paths worth backing up (e.g. ~/.config, custom scripts)
  • Common junk patterns to exclude (e.g. node_modules, *.log, cache/)

Shows suggestions and asks for confirmation before changing config.yaml. Credential stores such as ~/.ssh and ~/.gnupg are not auto-suggested. Add them only after explicit user approval and only when repository access and password-file backup are strongly controlled.

When the user asks to clean up or free disk space

sudo bash {baseDir}/bin/prune.sh

Options:

  • --keep-last 24 — keep only last 24 snapshots
  • --older-than 7d — remove snapshots older than 7 days
  • --dry-run — preview what would be removed
  • --yes — skip confirmation prompt

When the user asks to run a dry-run or test backup

Dry-run (validates without writing):

sudo bash {baseDir}/bin/backup.sh --dry-run

Self-test (full backup→restore→verify roundtrip in temp directory):

bash {baseDir}/bin/test.sh

Important notes

  • Silent by design: cron/systemd runs every hour at :05 and logs to the configured log file. No output unless there is a failure.
  • Telegram fires only on failure. If the user has not configured bot_token and chat_id, failures are logged only.
  • This is the time machine layer. It protects against "the agent broke something in the last 3 days." It is NOT a disaster recovery backup — that should be handled by an off-VM backup (e.g. restic to a remote server).
  • Password: The restic repository is AES-256 encrypted. The password file location is configured in config.yaml (chmod 600). Losing it means losing access to all snapshots.
  • Never commit secrets.env or .pass files to git. They are excluded via .gitignore.

When the user asks to uninstall or remove Time Clawshine

sudo bash {baseDir}/bin/uninstall.sh

This removes all system artifacts (systemd timer/service, cron, logrotate, binary, lock/marker files) but preserves the backup repository and password file.

To also delete all backup data (irreversible):

sudo bash {baseDir}/bin/uninstall.sh --purge

The source files in the skill directory are never touched — can re-install with sudo bin/setup.sh.

When the user asks to check for updates

Run the status dashboard which includes update info:

sudo bash {baseDir}/bin/status.sh

Or check manually:

openclaw skills update quick-backup-restore

Note: backup.sh automatically checks for updates once per day (if updates.check: true in config). It logs a warning when a new version is available but never updates automatically.

More in Data & APIs