← Back to library

Hand off one offline task before you sign out

Scotty Beam's evening habit asks an agent what useful work it could finish while you are away, then waits for you to pick. Our version caps suggestions at three, requires a concrete done artifact plus what login or decisions it still needs, and only runs after you choose.

  • OutcomeBefore you go offline, get up to three useful task options with a concrete done artifact named for each, pick one, and wake up to a review-ready deliverable the agent prepared after your choice.
  • SetupMinutes
  • LevelStarter
  • CreditScotty Beam (@ScottyBeamIO)
View original
Works with
Grok BotChatGPT agentClaude

Published Oct 23, 2026Updated Oct 23, 2026AI dry run on sample input: passed Oct 23, 2026Source verified Oct 23, 2026

When should you use it?

  • You leave work in progress overnight and want the agent to prepare something reviewable by morning
  • You do not want the agent starting tasks until you pick one
  • You can name current goals and which tools the agent may use

Not for: Letting the agent start work without your pick; Sending email, booking meetings or changing live systems overnight without approval; Replacing a morning inbox and calendar brief

What do you need, and what do you get?

You need

  • Your current goals or active projects in a short list
  • Which connected tools the agent may read (calendar, docs, repo, etc.)
  • Any hard limits (no outbound messages, no spends, no production deploys)

You get

  • Up to three task options, each naming the done artifact, missing logins and decisions only you can make
  • After you pick: one review-ready deliverable with sources or open questions labeled
  • No work started before your choice; no outbound actions unless you later approve them

How does it work, step by step?

  1. State goals and limitsTell the agent your active goals and what it must not do while you are offline.
  2. Get three optionsAsk for at most three tasks that each name the done artifact and what login or decision is still missing.
  3. Pick oneChoose a single option (or none). The agent waits until you choose.
  4. Review in the morningRead the deliverable, resolve open questions, and only then approve any follow-on sends.

The prompt

Our own prompt, inspired by Scotty Beam (@ScottyBeamIO) and written to work in Grok Bot, ChatGPT agent and Claude. Values in lime are placeholders.

Fill in the blanks (optional; the copied prompt uses your values)

Prompt3 variables
I am signing off for the day.
Active goals: {goals}.
Tools you may use: {allowed_tools}.
Hard limits: {hard_limits}.

Propose at most three useful tasks you could finish before I return. For each option list:
- Task name
- The exact artifact you would leave for me
- Logins or files you still need
- Choices only I can make

Do not start any task until I pick one option (or say none).
After I pick, complete only that task to a review-ready artifact. Label open questions. Do not send messages, spend money or change production systems unless a hard limit above explicitly allows it.

Inspired by Scotty Beam (@ScottyBeamIO): X Article: Grok Bot updates — evening offline handoff prompt. The prompt above is written by playbooks.bot in our own words; nothing from the original is quoted.

What can go wrong?

  • If the agent starts before you pick, stop and reset the rule.
  • Vague notes like make-progress do not count; demand a concrete artifact you can open.
  • Do not grant spend, deploy or outbound-message rights for an unattended overnight run.
  • This is not a morning brief; keep inbox triage for a separate routine.

Our take

Unattended agents earn trust when they wait for a pick and deliver a bounded artifact. Beam's evening handoff idea is useful because it separates suggestion from execution: list a few options with clear done states, pause, then run only what you chose. We keep the playbook to that habit and leave product pricing, email-claim setup and model menus out of scope. Distinct from a morning calendar-and-inbox brief, this is sign-off delegation. In our dry run we gave a fictional founder three active goals (pricing page copy, competitor note, contractor reply) and a no-outbound rule. The agent proposed three options that named done artifacts and missing logins, waited, then after we picked the competitor note produced a two-page comparison with source links and three unresolved questions marked for morning — and sent no email. Use it when goals are already clear; it will not invent priorities you never stated. Keep mornings for triage briefs if you need them — this playbook is only the sign-off handoff.

Where the time goes back: You stop returning to a blank slate; one chosen overnight task lands as something you can review instead of re-briefing from scratch.

FAQ

How is this different from ai-morning-brief-email-and-slack?
That playbook builds a morning digest from calendar, tasks and inbox. This one is an evening handoff: suggest options, wait for your pick, then prepare one review-ready artifact overnight.
Will it start all three tasks?
No. It suggests up to three and waits. Only the one you choose should run.
What counts as a finished output?
A concrete artifact you can review: a sourced brief, a draft doc, a comparison table or similar — not a vague status note.

Source and credit

Inspired by Scotty Beam (@ScottyBeamIO). We describe the idea in our own words, link to the original, and wrote the prompt ourselves; nothing is quoted.

Inspired by
X Article: Grok Bot updates — evening offline handoff prompt
Creator
Scotty Beam (@ScottyBeamIO)
Published on
X
Original date
Oct 8, 2026
License
Not openly licensed (the poster keeps all rights), so we quote nothing and link to the original
Originally for
Grok Bot
AI dry run
Oct 23, 2026: passed. Dry-run tested by our AI test agent on sample input. Waited for pick; review-ready artifact only; hard limits held.
AI dry run on sample data; not run with live connected accounts. The playbooks.bot test agent filled the prompt with synthetic inputs, produced the output and checked it against the page's stated outputs and guardrails.
Verified
Oct 23, 2026: source link resolves to the original, the credited creator published it, our excerpt or prompt matches it, and the license allows our use.

Are you the creator? Claim or correct this page.

More in Operations & Admin

All 203 Operations & Admin playbooks →

Related playbooks

Operations & AdminPrompt✓ AI dry run

Aim the agent at a janky internal app after it maps the UI

Have the agent study a janky internal UI out loud, complete one end-to-end task with a recording, and stop for login walls instead of guessing credentials.

Credit: Carnage (@0xCarnagee)

Under an hour
Operations & AdminPrompt✓ AI dry run

Brief an agent with nine fields, not a how-to

Before you hand a job to an agent, fill in nine short fields that name the result, the sources it may use, the deadline, what it may do alone, what needs your say-so, what it must never do, how to prove the work is finished, what to do when it gets stuck, and where to put the deliverable, then add a line that tells it to pick the simplest reliable method.

Credit: Michael Fenech (@Michael_Fenech_)

Minutes
Operations & AdminPrompt✓ AI dry run

Copy two working bot roles before you invent one

Start from a short list of narrow, already-used agent roles mapped to your week instead of a vague catch-all Assistant.

Credit: Carnage (@0xCarnagee)

Minutes
Operations & AdminWorkflow✓ AI dry run

Forward an email and let an assistant book the meeting

Forward a meeting request to a labeled inbox and have an AI assistant coordinate a time with the other person and add a video link.

Credit: Eddie Soehnel (@edsoehnel)

Under an hour