All use cases

Work

Back up the bot VM to R2 (or B2) every 15 minutes

Git + a reinstall script is not enough when a bot VM wipe also eats workspace files.

Share

What hurt

Sun of Dog (@thesunofdog) learned the hard way: git origin for code plus a setup script for installed software still left ~3.6 GB of agent state, transcripts, and generated files unprotected. A vendor VM wipe after a Grok Bot upgrade cost five days of work and ~$300 in recovery tokens. Tip with author-published steps. Not affiliated with xAI.

What to install or create

Ask the bots to pipe full box state to Cloudflare R2 on a deterministic 15-minute schedule (B2 as a second pipe). Keep git for code; R2 for the rest.

Bots

  • No new bot. Follow the first-run steps.

Plugins

  • None on day one. Follow the first-run steps.

Routines

  • None on the first run. Watch one manual pass first.

The prompt

This is a tip, not a pasteable system prompt. The author did not publish one. Follow the first-run steps.

First run

  1. 1Read https://x.com/thesunofdog/status/2106138803702472919 (credit @thesunofdog).
  2. 2Keep git for code; list the non-git paths that matter (agent state, transcripts, generated files).
  3. 3Ask your bots to set up a Cloudflare R2 (or Backblaze B2) backup of those paths — at ~3.6 GB the R2 free tier often covers it.
  4. 4Schedule a deterministic backup every 15 minutes that does not burn chat tokens.
  5. 5Do one restore drill on a spare folder before you trust the pipe; add a second B2 pipe later if you want redundancy.

Approval fence

Confirm what leaves the box (paths, secrets scrub) before the first live backup upload.

What to tweak

Steal the R2/B2 backup pipe. Free at a few GB, no tokens burned on the schedule.

Source

Sun of Dog @thesunofdog · Original post

Not affiliated with xAI. We did not write this prompt. Third-party template terms still apply.

Browse templates

Browse team blueprints

Pro pack

Want the full setup, not just the idea?

Pro packs give you the job text, prompts, schedules and pitfalls for popular use-cases.

See Pro