Grokbot TemplatesGrokbot Templates
All use cases

Engineering

Operating contract: role, outputs, verify, then automate

Jumping to a routine before the skill is proven just schedules failure.

Share

What hurt

@0xChaseTM shared Lauren Tan's agent formula via skill-to-automation: paste an operating contract (one bounded role; outputs = artifact + checks + blockers; observable verify; escalate on missing_access / unclear_scope), compare current vs revised skill on the same tasks, and only automate after repeated_success + clear_trigger. Author-quoted contract fields. Not affiliated with xAI.

What to install or create

Paste contract → revise skill → same-task A/B → automate only after repeated success.

Bots

  • Skill owner bot

    Runs one bounded role under the operating contract; escalates blockers.

Plugins

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

Files

  • Operating contract fields (role / outputs / verify / escalate_on).

Routines

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

The prompt

Skill operating contract (author-quoted fields)

Quoted from @0xChaseTM — as published. Do not invent a second system prompt.

Last line is our approval fence: Never send, post, spend, or buy without my explicit yes.

role: one bounded responsibility
outputs: artifact + checks + blockers
verify: observable acceptance criteria
escalate_on:
  - missing_access
  - unclear_scope

Keep preferences + skills persistent. Keep evidence + progress local to the task. Compare current_skill vs revised_skill on the same tasks, model and environment; record outcomes / failed_checks / human_corrections; keep the update only if observed execution improves. Automate only when repeated_success + clear_trigger. Define allowed actions and stop conditions before scheduling.

Never send, post, spend, or buy without my explicit yes.

First run

  1. 1Read https://x.com/0xChaseTM/status/2113766563727974465 (credit @0xChaseTM; Lauren Tan formula).
  2. 2Paste the author-quoted operating contract and fill one bounded role.
  3. 3Run the current skill, revise it, then retest on the same tasks/model/environment.
  4. 4Record outcomes, failed checks, and human corrections; keep the update only if execution improves.
  5. 5Schedule a routine only after repeated_success + clear_trigger; define allowed actions and stop conditions first.

Approval fence

Missing access or unclear scope stays supervised; you approve automation.

What to tweak

Steal the gate: repeated_success + clear_trigger before any schedule.

Source

Chase @0xChaseTM · 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